Simandou is set to add up to 120 Mtpa of high-grade iron ore once fully ramped, and the market has read that number as a straightforward reason for lower prices. This Insight asks a different question. What is the net change in global supply once the tonnes the industry is already losing every year, the years Simandou will take to reach full output, and the tonnes buyers actually want are all set against the headline figure. The strongest case against that reading gets equal space here too, including a major producer's own forecast that the higher-grade market barely grows before 2030. Samso Insights Metals & Mining Iron Ore Samso Market Strategy WHAT THIS PIECE COVERS 1.00 THE CONSENSUS EQUATION The bear case in one line, and whether the line holds. 2.00 A WORLD-CLASS DEPOSIT What Simandou is, stated before any of the qualifications. 3.00 NAMEPLATE IS A DESTINATION, NOT DAY ONE Where 120 Mtpa sits against where the project actually is. 4.00 THE TONNES THE WORLD IS ALREADY LOSING Depletion, the variable the consensus case leaves out. 5.00 BEYOND CHINA Why steel demand is not a single country's story. 6.00 THE QUALITY SHIFT Why not all iron ore competes in the same market. 7.00 THE BENCHMARK INVESTORS MIGHT BE WATCHING Why the 62% Fe price is not the only price worth following. 8.00 THREE GENERATIONS, ONE VERDICT Where all of this leaves the consensus equation. 9.00 REFERENCES AND SOURCES Every figure traced back to its primary source. A note on the vocabulary and the sources. Technical and mining terms are explained in plain English where they first appear, and all of them are collected in one box at the back. Every source carries an identifier, [R1] to [R31], and that identifier appears in the body text at the point the source is used, so any figure can be traced to where it came from. 1.00 — THE CONSENSUS EQUATION A single equation is carrying the whole argument - is this the Simandou factor The bear case on iron ore prices runs in one line. Simandou reaches its full 120 Mtpa, Vale keeps recovering toward its pre-2019 output, Australian exporters keep shipping near capacity, and Chinese steel demand keeps softening, so the seaborne market ends up longer than it is today and the price falls. Every part of that line is defensible taken alone. Set together, the equation only has one side filled in. The missing side is what happens to the tonnes the world already produces while the new tonnes arrive. Every iron ore mine has a finite life. Grades decline, strip ratios rise (the ratio of waste rock moved to ore recovered), and eventually the ore body is worked out or the cost of reaching what remains stops making economic sense. That process is running in the Pilbara, in Brazil, in Canada, in South Africa and in India at the same time as Simandou is being built, and it removes tonnes from the market every year whether or not a single new mine opens. The other missing piece is demand. China is the largest steel producer in the world and its consumption is slowing, but China is not the only buyer, and the growth happening in India, Indonesia, Vietnam and elsewhere does not show up in a story that treats "demand" as a synonym for "Chinese demand." This Insight does not argue Simandou is unimportant, or that prices cannot fall, or that the consensus view is wrong to watch the project closely. It argues that "more supply equals lower prices" only holds if nothing else in the ledger moves, and iron ore's ledger is already moving on the depletion side and the demand side in ways a single new mine, however large, does not fully offset. The six sections that follow test each part of the consensus equation in turn, starting with the deposit itself. It helps to place the argument inside a longer arc, sketched in FIG 02. For roughly the 2000s and early 2010s, iron ore competitiveness rewarded whoever could bring the most tonnes to market fastest, as China's construction boom absorbed almost any volume at almost any grade. Call that the first generation. From around the mid-2010s, competitiveness shifted toward cost. The Pilbara's low strip ratios and short rail hauls to port, and Brazil's high-grade ore bodies, turned Australia and Brazil into the world's lowest-cost seaborne suppliers, and the second generation rewarded whoever could land a tonne in China most cheaply. The question this Insight tests, discussion point by discussion point, is whether the industry is now moving into a third generation, in which the winning position is producing the highest-quality feedstock for a steel industry that increasingly needs it, rather than simply the most tonnes or the cheapest tonnes. Simandou fits uneasily across all three generations at once. It is being built with first-generation ambitions of scale, it will compete on second-generation cost once ramped given its grade and its rail-to-port logistics, and its ore quality gives it a genuine claim on the third generation this Insight argues the industry is moving toward. Whether that combination makes Simandou bearish for price, neutral, or something the market has not fully priced, is the question the rest of this piece works through. 2.00 — A WORLD-CLASS DEPOSIT Nothing that follows is intended to diminish it Simandou lies in the Simandou mountain range in south-eastern Guinea, a country most seaborne iron ore investors have never had reason to follow closely. It is worth being direct about scale before anything else. This is one of the largest untapped high-grade iron ore deposits in the world, and its arrival is a genuine event for the seaborne market, not a rounding error. The project is split into two halves. Blocks 3 and 4, in the south of the range, are held by SimFer, a joint venture in which Rio Tinto Group (ASX: RIO) holds 53% of the holding company Simfer Jersey Limited [R1], and the remaining 47% is held by Chalco Iron Ore Holdings, a consortium led by Chinalco and including Baowu, China Rail Construction Corporation and China Harbour Engineering Company. Simfer Jersey Limited in turn holds 85% of the project company Simfer SA, with Guinea's government holding the other 15%. Blocks 1 and 2, in the north, are held by Winning Consortium Simandou, in which Winning International Group and Weiqiao Aluminium held 51% and Baowu Resources 49% as at Rio Tinto's own November 2025 disclosure, with Guinea again holding 15% of the operating company. Rail and port infrastructure are jointly owned through a third entity, La Compagnie du TransGuinéen, split 42.5% each between Simfer and Winning Consortium Simandou, with Guinea's 15% carried free. Reported ownership has continued to move since that disclosure. Reuters and Mining.com have reported that Baowu subsequently increased its stake in Winning Consortium Simandou to a controlling 51%, with the entity now described in some coverage as the Baowu Winning Consortium Simandou. [R5] Samso has not been able to confirm that shift against a primary statement from either party, and it is treated here as reported, not established. The ore itself justifies the attention. Rio Tinto's own reserve statement puts SimFer's proved reserves at 66.4% Fe (the proportion of iron contained in a tonne of ore) and probable reserves at 65.0% Fe, an average of 65.3% Fe across the reserve base. For comparison, the seaborne benchmark price has historically referenced ore at 62% Fe, and the major Pilbara export blends now stand at 60.6% to just under 61% Fe following the specification downgrades discussed in section 4.00. Combined nameplate capacity across both halves of the project is up to 120 Mtpa (million tonnes per annum) once fully commissioned, of which SimFer's own share is around 60 Mtpa. China's steel industry has an obvious and long-stated interest in a large new high-grade source outside Australia and Brazil, and Chinese state-linked entities are present across both halves of the ownership structure. PLATE 01 Simandou operations, Guinea. Reproduced from Rio Tinto's public media library (operations imagery, undated). No specific press release is tied to this particular image. None of the analysis in the sections that follow is an argument that Simandou is a poor deposit, badly located, or badly owned. It is the opposite. The project's scale and grade are exactly why the market has treated it as such a significant bearish input, and exactly why it is worth asking whether "significant deposit" and "significant near-term price impact" are the same claim. 3.00 — NAMEPLATE IS A DESTINATION, NOT DAY ONE 120 Mtpa is where the project is going, not where it is A nameplate capacity, the maximum output a mine or plant is designed to sustain once fully built and running, is a design target, not a day-one result. Every large mine in the world has taken years to move from first ore to a sustained run at nameplate, and Simandou's own disclosures suggest it will be no exception. First ore left Guinea in December 2025, with the project's "start of operations" formally marked on 11 November 2025 as testing and commissioning began across the mine, rail and barge port infrastructure. [R1] Rio Tinto's own guidance for 2026, the project's first full year, is for total Simandou sales of just 5 to 10 million tonnes on a 100% basis, against the combined 120 Mtpa design capacity. [R3] That is low single-digit output as a share of nameplate in year one. SimFer has separately guided to a 30-month ramp-up from first ore to reach its own 60 Mtpa share [R2], implying full SimFer output is not expected before roughly mid-2028, with Mining Weekly separately reporting a similar "H2 2028" target for that half of the project. Analyst firm Breakwave Advisors described the project's near-term effect in January 2026 as "more symbolic than volumetric," estimating 2026 production at only 15 to 20 million tonnes. [R8] How unusual is a multi-year ramp-up. FIG 03 sets four comparable projects against Simandou's own guidance. Rio Tinto's Gudai-Darri mine in the Pilbara, a more straightforward open-pit project with existing regional infrastructure, still took roughly a year from first ore in June 2022 to its guided nameplate of 43 Mtpa. [R10] Hancock Prospecting's Roy Hill, also in the Pilbara, took closer to two years from first shipment in December 2015 to confirmed nameplate operation in September 2017. [R11] Anglo American's Minas-Rio in Brazil, which relies on a 529 kilometre slurry pipeline rather than rail, took longer again, and then lost roughly nine months of production entirely in 2018 to 2019 when the pipeline had to be shut down and inspected after a leak, meaning a durable nameplate run was still not clearly established years after first ore. [R9] Rio Tinto's own underground copper project at Oyu Tolgoi in Mongolia, included here as a risk comparable rather than an iron ore project, is the more cautionary case. Its original "early 2020" target for underground production slipped roughly three years to 2023, with disclosed cost overruns of at least US$1.9bn against an original US$5.3bn estimate and final capital cost above US$7bn, driven by shaft-sinking delays and more difficult ground conditions than expected. [R12] Simandou is not an underground mine, and the comparison is not a prediction that it will suffer the same fate. It is a reminder that even a major diversified miner with deep project-delivery experience can see a schedule slip by years once ground conditions, weather or logistics turn out harder than planned, and Simandou combines a new 600-plus kilometre rail corridor, a new barge port at Morebaya, and Guinea's own political and workforce risk in a way none of the Pilbara comparables had to manage. Simandou has already produced examples of exactly that kind of friction. A locomotive supply dispute in November 2025 delayed early shipments and stranded over 150,000 tonnes of ore. A fatal incident at the mine site in February 2026 led Rio Tinto to suspend SimFer's mine operations while it investigated. A nine-day strike across Baowu and Winning Consortium Simandou operations in April and May 2026 [R6], involving roughly 3,000 workers in a dispute over pay parity with the neighbouring SimFer operation, was resolved through a tripartite agreement with Guinea's government. As construction winds down from a peak workforce reportedly above 60,000 toward a planned operational headcount under 15,000, mass layoffs at individual work sites have drawn scrutiny from Guinea's own mines ministry. [R7] None of this makes Simandou unusual for a project of its size in a developing country. It does mean the 120 Mtpa figure is a multi-year destination the project is still travelling toward, not a switch the market can assume is already on. There is also a live discrepancy in how far along the project actually is. SimFer's own Q1 2026 update, dated 26 March 2026, reported the mine 73% complete and the port 75% complete. Mining Weekly, reporting in July 2026 on the same 31 March 2026 reference date, cited 74% and 78% completion [R4] respectively. Samso has not found a primary release reconciling the difference, and treats SimFer's own figure as the more reliable of the two while flagging the gap rather than picking a number and moving on. SAMSO ACCESS Some of the work does not appear here Samso holds back a set of Insights for members. Same method as this piece, the same sourcing and the same counterweights, on companies and sectors that do not make the public feed. ● Every member-only Samso Insight unlocked ● Full access to the gated Insights archive ● New member-only pieces included as they publish 4.00 — THE TONNES THE WORLD IS ALREADY LOSING Depletion is the variable the consensus case leaves out Every seaborne iron ore mine eventually runs out of easily accessible high-grade ore. Grades decline, strip ratios rise, and processing plants built for one ore specification have to be adjusted or the product specification quietly lowered. This is the variable missing from the simple consensus equation, and BHP (ASX: BHP), one of the two largest seaborne exporters in the world, has put a number on it. In its own Economic and Commodity Outlook, published in August 2025, BHP estimated that approximately 250 million tonnes of operating seaborne iron ore supply will deplete between 2025 and 2035. [R13] BHP separately identified around 180 million tonnes of existing supply produced at US$80 to 100 per tonne cost, and a further 110 to 120 million tonnes above US$90 per tonne, as marginal capacity vulnerable to closure if prices fall. This is not an independent analyst's estimate. It is BHP's own company forecast, published to explain its own investment case, and it should be read with that context in mind rather than as a neutral third party's number. Samso has deliberately not combined BHP's depletion estimate with Simandou's 120 Mtpa into a single net supply figure, shown separately in FIG 04. The two numbers describe different things, a cumulative ten year run-off against a single project's eventual annual rate, and turning them into one figure would create a precision the underlying data does not support. The depletion story is visible mine by mine, not just in one company forecast. In the Pilbara, BHP cut its Mining Area C Fines specification to 60.6% Fe from July 2024, and Rio Tinto has lowered its flagship Pilbara Blend Fines from a typical 61.6% Fe to below 61% Fe as it described its own "ore quality weakening" in a March 2025 statement. [R15] S&P Global Platts, which sets the industry's reference benchmark price, has itself been moving that benchmark from 62% Fe toward 61% Fe to reflect the sector-wide grade decline, and Fastmarkets launched a separate 61% Fe fines index in mid-2025. Mineral Resources closed its entire five-mine Yilgarn hub in Western Australia in mid-2024, citing limited remaining mine life, at a cost of around 1,000 jobs, and BHP announced in October 2025 that it would wind down its Yandi mine through 2026 as that operation's output continued its multi-year decline. [R16] Brazil tells a version of the same story from a different starting point. Vale produced 336.1 million tonnes of iron ore in 2025, its highest output since 2018, a result the company reported in growth terms. [R17] What that framing leaves out is that 2018 output, immediately before the Brumadinho tailings dam disaster in January 2019, was 384.6 million tonnes. Seven years after that disaster, Vale's 2025 production remains roughly 49 million tonnes, or around 13%, below its pre-disaster peak. In South Africa, Anglo American's Kumba Iron Ore reported in its own December 2025 reserve statement that both its Sishen and Kolomela mines have a reserve life of 16 years, to around 2041 [R18], and described Sishen's reserves as showing "annual depletion," down 15.2 million tonnes, or 2%, over the preceding year. In Canada, Rio Tinto's Iron Ore Company has faced repeated operational setbacks, including a 2023 production cut tied to extended plant downtime and wildfire recovery. India tells the same story from an unexpected direction. Most investors associate India with rising steel demand rather than falling ore exports, and both are true at once. Iron ore exports have fallen from an average of 72 million tonnes a year between 2005 and 2014 to just 28 million tonnes a year between 2015 and 2024 [R19], as rapidly growing domestic steel demand absorbs an increasing share of a production base that has itself nearly doubled. The tonnes have not disappeared from the ground. They have been redirected inward, which for the seaborne market that Simandou sells into amounts to much the same thing as depletion. Historically, major producers have rarely volunteered this framing in their own investor communications, but that may be starting to change. Matthew Holcz, Rio Tinto's iron ore chief executive, told a Melbourne Mining Club lunch on 5 August 2026 that "the rate of depletion is very much underestimated," according to a Reuters report carried by MINING.com the same day. [R14] He pointed to Pilbara assets built during the 2005, 2010 and 2015 investment waves that are now 15 to 20 years old and running at higher marginal cost. Rio Tinto has committed to spending more than US$13bn on new Pilbara mines, equipment and infrastructure between 2025 and 2027, and Holcz quantified the wider problem directly, estimating the industry needs to add 800 million tonnes of new supply globally over the coming decade just to hold output steady, against only 300 million tonnes currently committed. He also pushed back on recurring predictions of the sector's decline, saying "it feels like every year, the demise of iron ore is very much being exaggerated." A company executive making the public case for higher prices deserves the same caution Samso applies to BHP's own depletion estimate earlier in this section, since Rio Tinto benefits directly if the market believes supply is tighter than it looks. Even allowing for that, it is the clearest on-the-record confirmation yet that a senior figure inside one of the world's largest iron ore producers sees the same gap this Insight has been describing. That kind of public statement is still the exception rather than the rule. Rio Tinto's own December 2023 messaging on the Pilbara was built around expansion, a system capacity target of 345 to 360 million tonnes and a study into the Rhodes Ridge deposit, with no depletion language at all, even as the same period saw its product specifications quietly downgraded. The gap between the expansion story majors tell investors in routine disclosures and the grade and reserve-life data they also disclose is worth pausing on. It does not mean the majors are misleading anyone. Expansion and depletion can both be true of the same company at the same time, one describing total volume and the other describing the quality and duration of the ore body behind it. It does mean an investor reading only the headline production and capacity numbers is seeing one side of the picture. None of this amounts to a claim that global reserves are running out. The US Geological Survey's February 2026 Mineral Commodity Summaries put world iron ore reserves at 200 billion tonnes of crude ore containing 87 billion tonnes of iron, and world resources, a broader and less certain category than reserves, at more than 900 billion tonnes of ore. [R20] The point is narrower and more relevant to a pricing question. A large share of the capital being spent across the industry, Simandou included, is not simply adding to total supply so much as replacing tonnes the market is simultaneously losing elsewhere, and BHP's own 250 million tonne, ten-year depletion estimate is larger than Simandou's entire nameplate capacity even before the years Simandou will spend ramping up are taken into account. 5.00 — BEYOND CHINA Demand is not a single country's story China produces and consumes more steel than any other country by a wide margin, and its slowdown is real. Apparent steel use per capita, a country's steel consumption divided by its population and the standard measure economists use to compare how steel-intensive different economies are, fell from 634.9 kilograms in 2023 to 601.1 kilograms in 2024 in China, according to the World Steel Association. [R21] Chinese crude steel production fell from 1,005.1 million tonnes in 2024 to a provisional 960.8 million tonnes in 2025, and the industry's own Short Range Outlook forecasts a further 2.0% demand decline in 2025 and 1.0% in 2026. None of that is in dispute, and it is the single largest reason the consensus bear case exists. What the consensus case tends to skip is that "demand" is not a synonym for "Chinese demand," as FIG 05 sets out by country. India's apparent steel use per capita rose from 93.0 to 102.6 kilograms over the same year, and its crude steel output has climbed from 120.3 million tonnes in the 2021 to 2022 financial year to a record 169.2 million tonnes in the 2025 to 2026 financial year, an average annual growth rate of close to 9%. India's government has an official target of 300 million tonnes of steel capacity by the 2030 to 2031 financial year, and officials have floated an aspirational 500 million tonnes by 2047, though that longer-range figure is a stated ambition rather than codified policy. The World Steel Association's Short Range Outlook forecasts roughly 9% Indian demand growth in both 2025 and 2026. India is not alone. Indonesia's crude steel production rose 7.1% to 18.0 million tonnes in 2024. Vietnam's rose 14.6% to 22.0 million tonnes, the fastest growth rate among the major producers the World Steel Association tracks. The Middle East's apparent steel use per capita rose from 193.0 to 196.7 kilograms, with Saudi Arabia named alongside Vietnam, Egypt and India as a driver of the industry's expected 2026 demand rebound, even though Saudi Arabia's own domestic production actually fell slightly over the same period, a reminder that rising demand and rising domestic output are not the same thing when a country imports the difference. Africa's steel demand, flat for the best part of a decade at 35 to 40 million tonnes, grew at an average of 5.5% a year across 2023 to 2025 to reach around 41 million tonnes, against a continental infrastructure financing gap the African Development Bank put at US$75bn a year in a November 2025 report. [R22] None of these countries individually approaches China's scale, and none is likely to on its own. Set together, though, they describe a demand base that is broadening rather than simply shrinking, at exactly the moment the consensus supply case assumes demand is a single, softening number. There is a genuine measurement complication worth flagging here. The World Steel Association's own forecast for global 2026 demand growth was revised down between its October 2025 outlook, which put global growth at 1.3%, and an April 2026 update cited by India's Ministry of Steel, which put it at just 0.3%. Both figures come from the same forecaster months apart, and the gap between them is itself a sign of how much genuine uncertainty runs through the aggregate global number, even before it is broken down by country. There is a second demand channel, separate from population and income growth. Resource-producing developing economies have spent the past few years renegotiating for a larger share of the value their own resources generate, rather than simply exporting them under older contracts. In Malaysia's Sarawak state, a 5% sales tax the state now levies on petroleum products raised RM (Malaysian ringgit) 2.5bn to RM4.8bn a year, depending on oil prices, according to the state government's own figures, and its 2026 budget projects RM3.5bn from that tax alone. [R23] Sarawak has also legislated, under a 2023 ordinance, for its own state gas company to become the state's sole gas distributor, though that transition is still being contested with Petronas, the national oil company, and has not been fully resolved. Ghana has followed a similar pattern in mining. The government's free carried interest, an equity stake it holds in a project without paying for it, in the Ewoyaa lithium project was increased from the 10% set by national law to 13% under a renegotiated agreement, though the lease was still moving through parliamentary ratification as of November 2025. [R24] The premise worth testing is that revenue captured this way tends to fund the kind of domestic infrastructure, roads, ports, housing and power, that first-world economies built decades ago and that steel-intensive construction still requires. Sarawak's Premier, Abang Johari Openg, said in May 2026 that "stronger revenue means greater capacity to invest in public development." Samso has not found data tying either arrangement directly to steel demand, and the sums involved are small next to China's economy. Sarawak's population is around 2.8 million and Ghana's around 34 million, against more than a billion tonnes of Chinese steel consumption a year. The mechanism is plausible rather than proven. It is worth watching precisely because it lies outside the China-only framing the consensus case relies on, and because a growing number of resource-producing economies appear to be moving in the same direction. There is also a case that the effect works less directly than a government simply spending its own new revenue. Captured revenue can strengthen a government's own investment case, with agencies that look well funded and prudently run appearing a safer, more credible destination for other people's capital. Sarawak's Premier was reported in February 2024 as saying that prudent financial management of the state's finances, funded in large part by its growing petroleum revenue, would attract more investors to the state. Sarawak's own approved investment figures give some support to that reading. Total approved investment reached RM10.4bn in the first half of 2024, roughly evenly split between foreign and domestic capital, and the state posted a further RM4bn in the first quarter of 2026, the fifth-highest of any Malaysian state that quarter. [R25] The connection is not clean. Most of that 2026 figure, RM3.2bn of the RM4bn, was itself oil and gas exploration investment, the same sector generating the state's revenue, rather than evidence of a wider diversification effect. Analysts covering Ghana's lithium sector separately warn that whether resource revenue growth broadens into wider investor confidence depends on governance credibility and consistent policy implementation, not on the size of the government's take alone. [R26] The mechanism Samso would expect, revenue capture improving a government's fiscal standing and in turn its attractiveness to capital beyond the resource sector itself, is plausible and partly supported by Sarawak's own investment data. It is not yet demonstrated cleanly enough to treat as settled. COFFEE WITH SAMSO The same companies, on the record Coffee with Samso is the interview series. Management answering questions in long form, so the people behind a project can be judged alongside the numbers in a piece like this one. Free to watch. Companies commission an interview, and it stays free to everyone, permanently. Watch the interviews → 6.00 — THE QUALITY SHIFT Not all iron ore is competing in the same market The traditional route to steel, the blast furnace, smelts iron ore with coke, a processed form of coal, in a large furnace, and is responsible for the majority of the steel industry's carbon emissions. Two other routes are growing their share of global output, and both need a different kind of ore than the blast furnace does. The electric arc furnace, or EAF, melts scrap steel or direct reduced iron using electricity rather than coke, producing far less carbon per tonne of steel made. EAF's share of global crude steel production rose from around 26% in 2020 to 28.6% in 2023 and 29.1% in 2024, according to the World Steel Association, and Wood Mackenzie's October 2023 analysis projects that share reaching 50% by 2050. Direct reduced iron, or DRI, is made by removing oxygen from ore using heat and a reducing gas without fully melting it, and DRI feeds electric arc furnaces in place of scrap where scrap supply runs short. Global DRI production rose from 106.2 million tonnes in 2020 to 144.1 million tonnes in 2024 on the World Steel Association's count, or to 140.8 million tonnes on the independent Midrex count. [R27] The two series differ because they use different methodologies, and should be cited separately rather than averaged. The catch, and the reason this section is directly relevant to the Simandou debate, is that direct reduction needs ore purer than most of what the seaborne market currently supplies. Midrex, the industry's leading DRI technology licensor, specifies a minimum of 66% Fe, ideally above 67%, with low acid gangue content, to feed a direct reduction plant reliably. BHP itself has separately stated that DRI production needs ore in the range of 67% Fe. [R28] Simandou's average reserve grade of 65.3% Fe falls close to, but just below, that threshold on the figures Rio Tinto has disclosed, and no primary source Samso has reviewed confirms the ore ships as direct-reduction-ready without further beneficiation, the process of upgrading raw ore's iron content before it is sold. Simandou is a materially higher grade than the Pilbara benchmark, and it is a useful input for blending toward direct-reduction specification. It is not, on the disclosed numbers, automatically the same product as ore already refined to the 67% Fe threshold the technology prefers. Forecasts for how much extra DR-grade ore the market will need vary widely by source and should be treated as scenarios rather than facts. The International Iron Metallics Association estimated merchant DR-grade pellet demand rising from around 38 million tonnes in 2020 to 81 million tonnes by 2030. Wood Mackenzie identified 213 million tonnes a year of announced projects at 67% Fe or above, but rated only 41 million tonnes of that as "probable or highly probable." BloombergNEF's more aggressive scenario projects a DR-grade ore deficit of 15 million tonnes a year by 2030, widening to 133 million tonnes a year by 2040 under a net-zero pathway. [R29] These are modelled outcomes from named forecasters, not settled facts, and the range between them, from a modest gap to a large one, is itself informative about how uncertain the quality transition's pace really is. Not every major producer accepts that the quality shift will move fast enough to affect pricing. Vale, one of the two largest seaborne exporters, told investors in September 2021 that it expected no change in the proportion of seaborne ore above 66% Fe by 2030, with that share remaining at just 3% of the total market. [R30] If Vale is right, the high-grade segment of the market stays a small niche for years yet, direct reduction capacity grows more slowly than the more aggressive forecasts in this section suggest, and Simandou's grade advantage becomes less relevant to its pricing than its sheer volume. This is the single strongest challenge to the quality-led argument in this piece, and it comes from a company with every incentive to get its own market forecast right. 7.00 — THE BENCHMARK INVESTORS MIGHT BE WATCHING The 62% Fe price is not the only price worth watching Most reporting on iron ore prices quotes a single number, the benchmark price for fines at 62% Fe. That convention made sense when most seaborne ore traded close to that grade. It becomes less useful as a single reference point as the market splits into a lower-grade product competing on cost and a higher-grade product competing on quality, and an investor who only watches the 62% Fe benchmark may be watching one line on the chart rather than the line most relevant to a project like Simandou. Higher-grade ore already earns a premium over the benchmark, and lower-grade ore is sold at a discount to it. Fortescue (ASX: FMG), whose product grades below benchmark, realised only 73% and then 68% of the 62% Fe benchmark price across the first two quarters of its 2022 financial year [R31], an illustration of how material the discount for lower-grade ore can be in practice. Fastmarkets, the pricing agency, publishes both the 62% Fe benchmark and separate premium indices for higher-grade fines and pellets, confirming that a distinct higher-grade market exists and is priced separately. What Samso could not find, after searching public sources, is a published figure showing exactly how many dollars a tonne of premium each additional percentage point of Fe grade is worth. Those coefficients are held inside subscription-only Fastmarkets, Mysteel and Platts products and are not publicly disclosed in a form this piece can cite. Any investor relying on a specific per-point premium figure they have seen quoted informally should treat it as an estimate, not a published fact, unless they can trace it to one of those primary pricing services. Samso's own reading, after setting the forecasts and Vale's counter-view against each other, is that the fair position is uncertainty rather than conviction in either direction. The EAF and DRI growth trends in section 6.00 are real and moving in one direction, but the pace forecasters expect ranges from modest to dramatic, and Vale's flat 2030 estimate cannot be dismissed simply because it is the least exciting scenario on the table. What can be said with more confidence is that Simandou's grade gives it optionality a lower-grade Pilbara tonne does not have, whether or not the direct-reduction market grows quickly enough to reward that optionality at scale within this decade. SUPPORT SAMSO Independent, organic, and free to read We produce independent organic content and need your support to keep that going. Every contribution covers the cost of research, keeps the site running, and buys the time to write the next piece. A$10, A$20, A$30 or an amount you choose One time, weekly, monthly or yearly A donation buys nothing and unlocks nothing. It keeps the work going. 8.00 — THREE GENERATIONS, ONE VERDICT Where this leaves the consensus equation Return to FIG 02 and the three generations it sets out. The first generation rewarded scale, the second rewarded cost, and this Insight has argued the industry is moving into a third generation that rewards quality. Simandou is a genuine test case of that shift, a project scaled and financed like a first-generation mine, priced to compete like a second-generation one once ramped, and graded like a third-generation one from the start. Running back through the six discussion points, none of them individually overturns the consensus bear case, and this piece has not tried to make them. Simandou is exactly as large and as high-grade a deposit as its owners say it is. Its 120 Mtpa nameplate is real, even if it will take years and has already produced fatal incidents, strikes and logistics disputes on the way there. Mine depletion, at BHP's own estimated 250 million tonnes over ten years, is large enough to be significant but does not on its own cancel out a project of Simandou's scale. Demand outside China is growing but starts from a smaller base than China's own consumption. The quality shift toward direct reduction is real but its pace is disputed, including by Vale, one of the two companies best placed to know. Each variable, taken alone, leaves the consensus case largely intact. What changes when the variables are read together is the confidence the consensus case is entitled to claim. A single new mine, however large and however high-grade, is not entering a static market. It is entering a market that is simultaneously losing tonnes to depletion, gaining demand from countries the standard narrative rarely names, and slowly re-pricing quality relative to volume. None of those forces is large enough on its own to be a headline. Together, they are the reason "Simandou means lower prices" is a less certain conclusion than its confident repetition across 2025 and 2026 forecasting notes would suggest. The realistic position for an investor is not that Simandou is bullish or bearish. It is that the market may have done its supply-side arithmetic more carefully than its net-supply arithmetic, and the gap between the two is where the real research still needs to be done. 9.00 — REFERENCES AND SOURCES References & sources Figures below are drawn from the primary company, government and exchange sources named first in each entry, supplemented by the secondary press and analyst sources listed where no primary disclosure was found. Every company and analyst figure carries the date it was published or reported, and every figure that is Samso's own reading of an underlying disclosure is described as such in the body text rather than presented as fact. FIG 01 through FIG 06 are original Samso illustrations of the data and sources named in each caption. PLATE 01 through PLATE 03 are reproduced from Rio Tinto's public media library and are undated operational images not tied to a specific press release, as noted in each caption. [R1] Rio Tinto, "Simandou partners celebrate start of operations," 11 November 2025, and "Simandou iron ore project update," 6 December 2023, and "Conditions on Simandou investment now satisfied," 16 July 2024, riotinto.com. Ownership structure, nameplate capacity, ore grade, start-of-operations date. [R2] SimFer SA, "Rio Tinto SimFer provides update on the progress of the Simandou Project during the first quarter of 2026," 26 March 2026, and SimFer "About Us" page, simfer-sa.com. Completion percentages, ramp-up guidance, first ore and shipment dates, fatal incident confirmation. [R3] Rio Tinto, Q2 2026 results release, 15 July 2026. 2026 full-year Simandou sales guidance of 5 to 10 million tonnes. [R4] Mining Weekly, "Guinea iron-ore development reaches 74% completion," 24 July 2026, miningweekly.com. Alternative completion percentages for the same reference date, flagged in section 3.00 as unreconciled against the SimFer primary figure. [R5] Reuters and MINING.com, "China's Baowu takes control of Simandou iron ore operator." Reported shift in Winning Consortium Simandou ownership, not independently confirmed against a primary statement. [R6] Businesswire (Financial Content), 15 February 2026, and Discovery Alert, 7 May 2026, citing Reuters. Fatal contractor incident and April to May 2026 strike at Baowu and Winning Consortium Simandou operations. [R7] MINING.com, "Mass layoffs overshadow Guinea's Simandou mega-mine as output accelerates." Construction workforce wind-down and reported layoffs. [R8] Westpac (Justin Smirk), 19 December 2025. Goldman Sachs (Aurelia Waltham), 29 October 2025, reported via MINING.com. S&P Global Commodity Insights, 19 January 2026, citing CERA. Breakwave Advisors, 21 January 2026 and 27 May 2026, breakwaveadvisors.com. Analyst price and production forecasts, attributed as analyst views rather than fact. [R9] Anglo American, "Anglo American delivers first ore on ship from Minas-Rio," 27 October 2014, and "Anglo American suspends Minas-Rio iron ore operation for pipeline checks," 3 April 2018, angloamerican.com. Minas-Rio ramp-up and 2018 pipeline suspension. [R10] Rio Tinto, "Rio Tinto opens Gudai-Darri, its most technologically advanced mine," 21 June 2022, and "Rio Tinto to increase Gudai-Darri iron ore mine capacity," 18 October 2023, riotinto.com. Gudai-Darri ramp-up and later expansion. [R11] Business News (WA), "Roy Hill hits 55mtpa rate," 4 October 2017, businessnews.com.au. Roy Hill nameplate confirmation. [R12] Rio Tinto, "Oyu Tolgoi partners reach comprehensive agreement...," 25 January 2022, and "Underground production celebrated at Oyu Tolgoi," 13 March 2023, riotinto.com. MINING.com coverage of the 2018 and 2019 delays and cost overruns. Oyu Tolgoi schedule and cost comparison. [R13] BHP, "Economic and Commodity Outlook," 19 August 2025, bhp.com. The 250 million tonne, 2025 to 2035 seaborne supply depletion estimate. [R14] Reuters and MINING.com, "Depletion of iron ore mines will underpin next decade's prices, Rio Tinto executive says," 5 August 2026, mining.com. Matthew Holcz's Melbourne Mining Club comments, the US$13bn Pilbara investment figure, and the 800 million tonne versus 300 million tonne global supply gap. [R15] S&P Global Commodity Insights, "BHP's iron ore brand specification revisions...," 5 June 2024, and "Australia's Rio Tinto to lower iron content of flagship Pilbara Blend Fines...," 28 March 2025, spglobal.com. Pilbara grade downgrades. [R16] Mining Weekly, "MinRes to shut Yilgarn iron-ore mine," 19 June 2024. Resources Review, "BHP to scale down operations at Yandi mine," 29 October 2025. Pilbara mine closures and wind-downs. [R17] Mining Weekly, "Vale's iron-ore output hits seven-year high in 2025...," 28 January 2026, reporting Vale's own Q4/FY2025 production report. Vale 2025 and 2018 production figures. [R18] Anglo American Kumba Iron Ore, "Ore Reserve (and Saleable Product) and Mineral Resource Report 2025," 31 December 2025, angloamericankumba.com. Sishen and Kolomela reserve life and depletion language. [R19] CRU Group, "All you need to know about Indian iron ore supply," 2025, crugroup.com. Indian iron ore export decline. [R20] US Geological Survey, Mineral Commodity Summaries 2026, February 2026, and Mineral Commodity Summaries 2025, January 2025, pubs.usgs.gov. World iron ore reserve and resource estimates. [R21] World Steel Association, World Steel in Figures 2025, and Short Range Outlook, 13 October 2025, worldsteel.org. India Ministry of Steel, "An Overview of Steel Sector," April 2026, steel.gov.in. Steel use per capita, production and demand forecasts by country. [R22] African Development Bank, 2025 Africa Investment Forum panel report, 28 November 2025, afdb.org. African infrastructure financing gap. [R23] The Star, "Higher oil prices to buoy Sarawak's state sales tax revenue," 26 May 2026, thestar.com.my. Malay Mail, "Petros now sole gas distributor as Sarawak legal advisor says state received RM49b in royalties, Petronas kept the rest," 4 September 2025, malaymail.com. Sarawak's petroleum sales tax revenue, the Premier's quoted remarks, and the contested status of Petros as gas distributor. [R24] Ecofin Agency / FinancialContent, "Ghana's Ewoyaa Mining Lease Ratification Advances: A New Dawn for African Lithium Supply," 12 November 2025. Natural Resource Governance Institute (Thomas Scurfield and Denis Gyeyir), "Assessing the Fiscal Regime in Ghana's Lithium Agreement," 22 July 2024, resourcegovernance.org. Ghana's free carried interest in the Ewoyaa lithium project and its parliamentary ratification status. [R25] The Borneo Post, reported via PressReader, 16 February 2024. Sarawak Premier's reported remarks connecting prudent financial management to investor attraction. Malay Mail, "Awang Tengah: Sarawak Records Total Approved Investment of RM10.4b in H1 2024," 29 July 2024, malaymail.com. DayakDaily, "Sarawak Draws RM4 bil in Approved Investments for Q1 2026," 2026, dayakdaily.com. Sarawak approved investment figures and sector breakdown. [R26] Bloomsbury Intelligence and Security Institute, "Ghana's Lithium Strategy: Can Value Addition Happen Without Deterring Investment?," 17 March 2026, bisi.org.uk. Governance credibility and policy implementation as conditions for investor confidence in Ghana's mining sector. [R27] World Steel Association, World Steel in Figures 2025, worldsteel.org, and Midrex Technologies, "World DRI Production Reaches 140.8 Mt in 2024," 25 August 2025, midrex.com. EAF share and DRI production figures. [R28] Midrex Technologies, "The Iron Ore Challenge for Direct Reduction on Road to Carbon-Neutral Steelmaking," June 2022, midrex.com. BHP statement on DRI ore quality, October 2021, reproduced via IEEFA, "Iron Ore Quality a Potential Headwind to Green Steelmaking," June 2022. DR-grade Fe thresholds. [R29] Wood Mackenzie, "Steel decarbonisation to redefine supply chains by 2050," 19 October 2023, woodmac.com. International Iron Metallics Association and BloombergNEF estimates, reproduced via IEEFA, June 2022. DR-grade ore demand forecasts. [R30] Vale, September 2021, reproduced via IEEFA, "Iron Ore Quality a Potential Headwind to Green Steelmaking," June 2022. Vale's own forecast on the share of seaborne ore above 66% Fe by 2030. [R31] Fastmarkets, "How Is Iron Ore Priced? VIU & Index Explainer," updated 14 July 2026, fastmarkets.com. Fortescue financial year 2022 realised pricing data, reproduced via IEEFA, June 2022. Benchmark pricing methodology and grade discount example. Cover photo: Hematite (GeoDIL number - 1313), Darla Sondrol. CC0, via Wikimedia Commons (https://commons.wikimedia.org/wiki/File:Hematite_(GeoDIL_number_-_1313).jpg). Reproduced photograph, shown faded under a navy overlay on the article thumbnail. The Samso Way – Seek the Research Here at Samso, we pride ourselves on delivering content for investors that is independent and informed by over three decades of experience in the industry. Our content is well-researched and is only created if I see merit in discussing the company's story. Our mission is simple. Cut through the noise and spotlight what matters. Genuine stories, grounded insights, and real opportunity. Investors can explore our four core platforms: Coffee with Samso Samso Insights Samso News Samso Research There may be numerous paths to success in investing, but the common thread among successful individuals is that they remain committed to making informed decisions. Equip yourself with the right knowledge and tools, and you will be well on your way to achieving your financial goals. Most importantly, investors need to be absolutely diligent in understanding their own risk-reward tolerance and capabilities. Never bite off more than you can chew. As they say, Rome wasn't built in a day, and the Great Wall stood because it took centuries to complete. The Samso Philosophy: Stay curious. Stay sharp. And remember, digging deeper always uncovers the real value. In Life, there is no such thing as a Free Lunch. Never bite off more than you can chew is my parting comment. Happy Investing, and the only four-letter word you need to know is DYOR. To support our independent work, please head over to our Support Page and give us a helping hand in any of the ways listed. This is a new initiative for the Samso Platform, and it was always the concept of Samso when we started this journey in 2018. Disclaimer This article is general information only. It is not financial product advice and it does not take into account any reader's objectives, financial situation or needs. It contains no price targets, no valuations and no recommendation to buy, hold or sell any security. Companies named in this article are used as historical illustrations of events that have already occurred and no view is expressed on their current merits. Readers should consider obtaining advice from a licensed financial adviser and should conduct their own research before making any investment decision. Samso and its associates may hold interests in companies mentioned. Read full Disclaimer. About Samso Samso is a trusted platform that equips dedicated investors with up-to-date industry knowledge and insights from top CEOs and thought leaders. By staying informed on business advancements and market trends, investors can enhance their financial decisions through a combination of expert guidance and their own research. The Investor's Playbook is an open-ended, evergreen educational series about how to think about investing rather than what to buy. Samso Insights | www.samso.com.au | An Investor Lens on ASX-Listed Companies

Austral Resources Australia (ASX: AR1) spent more than two years suspended from trading before a recapitalisation, a $40 million placement and a $65 million follow-on raise put it back in business. It now controls two processing plants, has agreed to buy a former Glencore zinc mine for the copper ground underneath it, and has just reported a maiden drill result its chairman calls uncommon. None of that yet carries a JORC resource number, and the company is still recruiting for a permanent chief executive. This piece sets out what Austral has actually bought, what still needs to be proven, and what would have to go right for the number a broker has already put on the stock to hold up. Samso Insights Metals & Mining Copper Austral Resources WHAT THIS PIECE COVERS 1.00 FROM SUSPENSION TO RECAPITALISATION Two years off the boards, and what it took to come back. 2.00 MOUNT KELLY AND ROCKLANDS, THE TWO-HUB PLAN The two Queensland processing sites the operating model rests on. 3.00 LADY LORETTA, BUYING A ZINC MINE FOR ITS COPPER GROUND A 1969 zinc discovery bought for what sits around it. 4.00 WHAT THE $65 MILLION RAISE ACTUALLY FUNDS The two tranches, and where the money is going. 5.00 THE HAMMER METALS DEAL, A THIRD ASSET AND A LOT MORE DILUTION The August 2026 proposal, and what it costs existing holders. 6.00 CANYON, A MAIDEN INTERCEPT, NOT A RESOURCE What the June 2026 drilling found, and what it does not yet amount to. 7.00 WHAT THE MARKET IS ALREADY PRICING IN The bull case as others put it, including Rio Tinto's own chief executive. 8.00 WHAT THE SHARE PRICE ACTUALLY DID, EVENT BY EVENT Each announcement lined up against what the shares did next. 9.00 THE UNFILLED SEAT AND THE OTHER OPEN RISKS A vacancy at the top, and the risks that go with it. 10.00 REFERENCES & SOURCES Every dated figure traced to its announcement. A note on the vocabulary and the sources. Technical and mining terms are explained in plain English where they first appear, and all of them are collected in one box at the back. Every source carries an identifier, [R1] to [R31], and that identifier appears in the body text at the point the source is used, so any figure can be traced to where it came from.

SAMSO INSIGHTS · ARE WE LOOKING AT THE NEXT PGE PRODUCER? What a platinum group element project needs before it gets built, and what Bengwenyama's rock is actually worth. Southern Palladium has spent four years turning a shallow reef on South Africa's Eastern Limb into a mining right, an ore reserve and a metallurgical result that has changed what the project is worth. Chalice Mining has spent six years and much more money working towards something similar in Western Australia. Terra Metals has a discovery the market likes and no resource behind it yet. This piece asks one question. What does a platinum group element project need before it gets built, and how many of those things does Bengwenyama have? Samso Insights Platinum Group Elements ASX Small Caps Investor Awareness WHAT THIS PIECE COVERS 1.00 WHY I AM WRITING THIS One post about Southern Palladium, and a question it raised that I could not answer. 2.00 WHERE THE PROJECT IS Bengwenyama, nine kilometres west of Steelpoort, on the red band. 3.00 GROUND THAT HAS BEEN STUDIED BEFORE A dome named in 1995, and three papers older than the company that owns it. 4.00 WHAT A PGE PROJECT NEEDS BEFORE IT GETS BUILT The eight tests a project has to pass, and where most projects fail. 5.00 THE FIRST TEST, THE GRADE What Mintek measured in 1982, and what the grade means here. 6.00 THE E'S, AND WHICH ONES MAKE THE WORLD GO ROUND What the letters in 4E and 7E count, and which metals carry the value. 7.00 THE SECOND TEST, AN OREBODY THAT CAN BE MINED Seventy-three centimetres of reef, and everything that interrupts it. 8.00 THE THIRD TEST, WHAT IT COSTS TO BUILD What it costs to buy an ounce a year, compared across projects. 9.00 WHAT WE KNOW SO FAR, AND WHAT IS LEFT Three tests answered, five still open, and where Part 2 picks up. 1.00 — WHY I AM WRITING THIS One post, and a question I could not answer I came across a post from Bridge Street Capital Partners[R12] about Southern Palladium, and it started me thinking that this would be a good project to work through properly. I read through the Bridge Street work, and I felt that this is a great start for me to understand the PGE sector. I worked over three decades in the mineral resource industry, and have done Coffee with Samso interviews with companies in the PGE sector. However, I'm still not much more educated in what we should be looking for as investors, or what kind of boxes need to be ticked. I saw this as a challenge and an opportunity at the same time. To go properly into what Bridge Street have done with Southern Palladium, and try to get an understanding of the sector. History shows that Bridge Street are probably the most intimate with this project. They were corporate advisers to Southern Palladium, and lead manager on the company's initial public offering and on its subsequent capital raisings. The analyst who writes their research also holds shares in the company, which Bridge Street disclose themselves. Being that invested in a company's story is a good sign in one sense, because nobody knows the project better. Others will view it as non-independent. Both readings are fair. In Part 2 of this series we look properly at the research Bridge Street has produced. My search has shown that only two firms publish research on Southern Palladium and both have a commercial relationship with the company. This Samso Insight would give the Samso community an arm's length view on the potential of the project. What interested me was what makes this company tick. Why is it heading into production when so many others are not? What are the boxes that have to be ticked, and which ones should an investor be looking at in the first place? Looking around the internet to find examples of PGE stories, I realised that there are not many examples for the retail investing community to look at a platinum group element project closely enough to answer questions. If anyone have looked at a PGE project, for me, the difficulty is in the sheer volume of information a project to understand. For example, is it 3E? Is 4E better? What about 6E and 7E? And then, what percentage of each of those E's is the one that makes the world go around? Then you start looking at the depth and the length of intersections? What is good and What is spectacular? As seasoned ionvestors in the mienral resoruce sector, we can "assume" most parts but I think, new players will need the letters to be clekared up now, because they run through every number in this piece. Fig. 01 sets the four bases side by side. Each one adds metals to the one above it, which is why the same rock reports 8.12 grams per tonne on a 4E basis and 9.78 on a 7E basis. 3E counts platinum, palladium and gold. 4E adds rhodium. 6E adds ruthenium and iridium on top of that. 7E adds osmium as well. So the same rock reports a bigger number as more metals are counted. Southern Palladium's UG2 resource comes in at 8.12 grams per tonne on a 4E basis and 9.78 grams per tonne on a 7E basis.[R1] Same rock, same drilling, two figures nearly 20 per cent apart, and both correct. That matters when you put two projects side by side. Chalice's Gonneville is reported on a 3E basis. Bengwenyama is usually reported on 4E or 7E. Comparing one company's 7E number with another's 3E number tells you very little. The second half of the question, which percentage is doing the work, is the more interesting one. Section 6.00 gets to it. The short version is that on this reef platinum and palladium are close to even, and rhodium is a small number in the ground doing a large amount of the earning. 2.00 — WHERE THE PROJECT IS Nine kilometres west of Steelpoort, on the red band The Bengwenyama project is in the Greater Tubatse and Sekhukhune District Municipalities of Limpopo Province, in the northeast of South Africa. It covers two farms, Nooitverwacht 324 KT and Eerstegeluk 327 KT. The Mining Right granted in August 2026 runs over 5,280 hectares of them.[R4] Steelpoort is about 9 km east. Burgersfort is 20 km northeast. Pretoria is 250 km west-southwest. This is a working mining district with sealed roads, grid power, a water pipeline and a local workforce that has been mining for generations. Plate 01 places the project on the national map. Plate 02 is more useful, because it shows where Bengwenyama is within the Bushveld Complex. The Bushveld Complex is the largest layered igneous complex in the world. It formed between 2.06 billion and 2.058 billion years ago, when successive pulses of molten rock were injected into the crust and then cooled slowly, settling out into layers by density in much the way sediment settles in still water. Those layers can be traced for hundreds of kilometres. It outcrops in three main arcuate limbs, Western, Eastern and Northern. Bengwenyama is on the Eastern Limb. The layered sequence divides into zones, and Plate 02 colours them. The red band is the Critical Zone, and that is the one that matters. It carries both the Merensky Reef and the UG2 Reef, the two layers that hold most of the world's platinum group metals. The tan Upper Zone and the grey Main Zone above and below it carry no economic platinum. The project area is on that red band, immediately south of Steelpoort, in a belt that has been mined for platinum and chrome for close to a century. Modikwa is immediately north, owned jointly by Valterra Platinum and African Rainbow Minerals. Two Rivers is to the south, held by African Rainbow Minerals and Implats, and beyond it are Mototolo and Booysendal. Marula, an Implats operation, is further north. All of them mine the same reef Southern Palladium is proposing to mine. 3.00 — GROUND THAT HAS BEEN STUDIED BEFORE A dome named in 1995, and three papers older than the company - historical PGE hints This ground has been studied before, and the work was not done by the company that owns it now. In 1995, Roger Scoon of Trojan Exploration and B. Teigler of Rhodes University published a paper in Economic Geology called “A New LG-6 Chromite Reserve at Eerste Geluk in the Boundary Zone between the Central and Southern Sectors of the Eastern Bushveld Complex”.[R9] It reports an exploration programme Trojan ran on the farm between 1990 and 1993. In that paper, Scoon and Teigler write that structural complications in the Critical zone are attributed to a hitherto unrecognised fold, the Eerste Geluk dome. That dome is the structure that has removed the reef from part of Bengwenyama. It was identified, named after the farm, and published in a peer-reviewed journal thirty-one years before Southern Palladium's drilling confirmed it was bigger than expected. Ten years earlier, in 1985, S.B. Gain of the South African Development Trust Corporation published a detailed description of the UG-2 reef on the farm Maandagshoek, about 20 km northwest of Steelpoort, based on 22 boreholes.[R8] Three years before that, in 1982, Christian McLaren of the Council for Mineral Technology and Johan de Villiers of Rand Afrikaans University published a study of the platinum group chemistry and mineralogy of the UG-2 across the whole Bushveld. They used 23 borehole cores and identified more than six thousand individual mineral grains. Their stated purpose was to work out how to recover the metal.[R7] Earlier again, in 1929, Percy Wagner recorded old platinum workings on Eerste Geluk itself.[R14] The historic drilling was left out of the resource The Pre-Feasibility Study[R1] records that Trojan's drilling from 1990 to 1993 was not used in the current Mineral Resource estimate, because the drillhole data was incomplete or unavailable. The published work and the current resource model are separate exercises on the same ground. Watch the interviews → 4.00 — WHAT A PGE PROJECT NEEDS BEFORE IT GETS BUILT Eight things a project needs, and where most of them fail One of the most intriguing things about the platinum group element sector is how hard it is to say what a project actually needs before it can go into production. When we talk about being in production, what we are really talking about is whether something is economical or not. This is one I have struggled with over the years. What is the grade? What is the size? What is the combination that turns a PGE deposit into an economical PGE deposit? If there is any investor out there, whether you have mining or mineral exploration experience or not, I would challenge you that this is a hard one to crack. Compare it with the Bushveld Complex, where the mineralisation is measured in hundreds of kilometres of strike. On the ASX I do not think we have seen anything like that yet. However, time has changed. That is why this project intrigues me, and why I have been trying to get my head around what makes it tick. Hopefully by doing that, we can then make our own assessment of all the up-and-coming ASX companies playing in this sector. Several ASX companies have been working towards platinum group element production for years. Chalice Mining found Julimar in 2020 and its share price closed as high as A$10.17 in November 2021. Terra Metals reported 52.97 g/t over one metre at its Southwest prospect in January 2026, and its share price rose about 370 per cent over the twelve months to July 2026. Neither company is in production, and neither is near it. The time it takes for a mineral resource project to get to the mining stage is usually a long one, and is often measured in decades. So it is very normal for investors to see that projects take a long time to drill out and to establish a scoping study and a pre-feasibility study. That normally takes a lot more money and a lot more time than short-term investors have. PGE projects may take a little bit longer, due to the complexity of the myriad of products, their value, and their market valuation at the time. That brings us back to whether you are talking about 3E, 4E or 7E. I have never been involved in the downstream end of a PGE project. I'm not sure whether the complexities of a combination of so many minerals make things harder, or whether it is pretty much the same as a one metal project. Fig. 02 shows both. Chalice ran from about A$0.25 to a closing high of A$10.17 and back to A$1.38. Terra Metals moved on a single drill result. Neither line represents a producing mine. Below are the eight things a PGE project needs before it becomes a mine. These are not a Samso invention. Each one comes from something that has actually stopped a project, and the evidence for each is named. One. Grade high enough to justify a small plant A higher grade does not make the rock richer in any way that matters on its own. What it does is let you move less dirt to get the same amount of metal, and that allows a smaller plant. Section 5.00 goes into why that is where the value actually comes from. Southern Palladium's Ore Reserve[R1] is 31.72 million tonnes at 6.17 grams per tonne of 6E. Counting only the three metals Chalice counts, platinum, palladium and gold, the same reserve carries 4.74 grams per tonne of 3E. Chalice's Ore Reserve at Gonneville[R5] is 260 million tonnes at 0.86 grams per tonne of 3E. On the same basis, then, Bengwenyama's reserve grade is roughly five and a half times Gonneville's. The 3E figure for Southern Palladium is Samso's arithmetic from the company's own published prill split, so that the two can be read against each other rather than on different bases. That difference is what the plant sizes reflect. Southern Palladium's Optimised Pre-Feasibility Study[R2] proposes 1.2 million tonnes a year in Stage 1 for over 200,000 ounces of platinum group metals. Chalice's Stage 1 treats 5 million tonnes for about 150,000 ounces of 3E. A smaller plant needs a smaller mill, less power, less water, a smaller tailings facility and fewer people to run it. All of that costs less to build and less to operate. Two. An orebody that can be mined Width, depth, continuity and structure decide what can actually be extracted and at what cost. A drill core measures the rock it passes through. What comes out of a stope is a different number, because mining brings waste up with the ore, loses ground to structures and works around things the drilling did not see. One of the biggest structural damages that can be done to a perfectly designed mining operation is a change in one of those parameters. A slippage in a wall. A blast going wrong. Structural damage in an underground drive that forces a change to the whole operating system. The damage that turns a perfectly good mine into one you should not have touched can happen very fast. A recent example is Meeka Metals (ASX: MEK), a gold producer at the Murchison Gold Project in Western Australia, which poured first gold in July 2025.[R15] Production fell from 9,174 ounces in the December 2025 quarter to 6,083 ounces in the March 2026 quarter, with weather among the factors. The June 2026 quarter recovered only to 6,424 ounces, and the company reported that open pit mining continued to be affected by lower-than-expected contractor productivity, which delayed access to high-grade open pit ore. The response was to change the plan. Meeka is ending open pit mining in July 2026, preserving what is left of the open pit resource in the ground, and moving to underground ore, which it expects to make up 40 per cent of the mill blend in the September quarter. None of that is a geological problem. The resource is unchanged. What changed was access to the higher-grade material the plan was built around, first because of weather and then because of contractor performance, and the answer was to abandon one mining method a year into production. The share price more than halved over the same period. The production shortfall was not the only reason, since sentiment towards small-cap gold developers cooled at the same time and about 117.8 million shares were issued to acquire the Mt Holland South deposits. But it shows how quickly the parameters can move on a mine that was designed properly. Bengwenyama has real difficulties here. Section 7.00 sets them out in full. Three. Capital intensity a small company can carry The clearest way to compare any mining project, such as this PGE project, across different countries is capital cost per annual ounce of production. It removes the effect of scale and asks what it costs to buy an ounce a year. Southern Palladium's Optimised Pre-Feasibility Study puts Stage 1 development capital at US$219 million for over 200,000 ounces a year. That works out at roughly US$1,100 per annual ounce. Chalice's Pre-Feasibility Study, released on 8 December 2025[R5], puts Stage 1 development capital at A$820 million for about 150,000 ounces of 3E a year, and Stage 2 expansion capital at a further A$840 million. Stage 1 works out at roughly A$5,500 per annual ounce, or about US$3,900 at the exchange rate of 0.713 on 17 August 2026. Those calculations are Samso's, using each company's own published capital and production figures. Capital intensity is a property of the orebody. It only changes if the geology or the engineering changes. One thing has to be said alongside that, or the comparison is unfair. Gonneville also produces nickel, copper and cobalt, and Chalice's all-in sustaining cost is US$370 an ounce of 3E after those by-product credits, falling to US$50 an ounce in the first three years. Southern Palladium's all-in sustaining cost is US$969 an ounce of 6E in Stage 1 and US$821 across both stages. So Bengwenyama is much cheaper to build per ounce of annual capacity. Gonneville is much cheaper to run per ounce produced. Chalice describes its project as the lowest cost PGM mine in the western world and the lowest cost of any undeveloped project. These are different kinds of asset, and the tests do not all point the same way. Four. A conventional flowsheet New processing technology adds risk to both the schedule and the budget. Southern Palladium states its own position directly. In the metallurgical results released on 9 July 2026[R3], the company writes that no novel technology is being applied, and that existing technologies are being optimised for the Bengwenyama orebody. Chalice shows what the alternative can cost. Its Pre-Feasibility Study announcement records an investment of approximately A$15 million into metallurgical test work and flowsheet design. That work generated over 100 samples from 33 dedicated metallurgical drill holes and involved more than 1,400 metallurgical tests, running through multiple phases and flowsheet iterations since testing began for the Scoping Study in 2021. Four years and A$15 million to settle a flowsheet, on a project that has not yet been built. Five. A concentrate someone will buy More projects are stopped by this than by any of the other seven, and it gets talked about the least. A mine sells concentrate to a smelter on agreed terms. If no smelter wants that concentrate, or the ones that do are already full, the project does not proceed no matter how good the orebody is. Bridge Street's February 2025 report[R12] records that Ivanplats secured smelting capacity with Northam, and that this appears to have locked Platinum Group Metals' Waterberg project out of the South African market for the time being. Waterberg holds 18.7 million ounces. That has not been enough to secure it a route to market. Six. A reserve rather than a resource A Mineral Resource says the metal is in the ground. An Ore Reserve says someone has tested whether it can be extracted at a profit, applying the mining, metallurgical, economic, legal and environmental factors that decide it. Southern Palladium declared a maiden Probable Ore Reserve of 6.29 million ounces in October 2024. Chalice declared a maiden Ore Reserve of 7.1 million ounces of 3E in December 2025.[R6] Both companies have passed this test. Terra Metals has not. Its Southwest prospect, which is what the market is pricing, has no resource at all, and the company's own JORC table states that the drill spacing is not currently sufficient to establish continuity. Seven. Permits In South Africa, mineral resources are held by the state as custodian. A company does not buy the right to mine. It applies for a grant of that right under the Mineral and Petroleum Resources Development Act. Southern Palladium applied in September 2023 and was granted its Mining Right on 7 August 2026.[R4] That wait is a big part of why Bengwenyama has stayed undeveloped for so long, even though it is one of the last shallow outcropping sections of the UG2 on the Eastern Limb. Eight. Being able to raise the money on the day The eighth item works differently from the other seven, because it is not a property of the project. Southern Palladium's peak funding requirement is US$279 million. Whether that can be raised depends on the company's market capitalisation and what the market is willing to fund at the time. Both of those move without anything at Bengwenyama changing. Southern Palladium's own share price shows this. It has ranged from A$0.195 in May 2025 to A$2.71 in February 2026, and closed at A$1.93 on 17 August 2026. On the 132.3 million shares on issue at 1 June 2026, that last figure is a market capitalisation of about A$255 million. The orebody did not change across any of it. Fig. 03 puts the company's own milestones against that line. The Pre-Feasibility Study, the Optimised Pre-Feasibility Study, the metallurgical results and the grant of the Mining Right are all marked, and none of them accounts for the size of the swings between them. The first seven items describe the project. The eighth describes market conditions, and those change. 5.00 — THE FIRST TEST, THE GRADE What Mintek measured in 1982, and what it means here Southern Palladium reports the UG2 Mineral Resource at 8.12 grams per tonne on a 4E basis and 9.78 grams per tonne on a 7E basis. Section 1.00 sets out what those letters count. The basis matters here, because 8.12 and 9.78 describe the same rock. What an independent study found In 1982, Christian McLaren of the Council for Mineral Technology and Johan de Villiers of Rand Afrikaans University sampled the UG-2 across the whole Bushveld Complex. They used 23 borehole cores and their stated purpose was to work out how to recover the metal.[R7] They divided their results by sector, and Table 01 sets out what they found. Bengwenyama is in the central sector of the Eastern Limb. On McLaren and de Villiers' sampling, that sector runs at 10.05 parts per million against a Bushveld-wide UG-2 average of 6.93. The company reports 9.78 grams per tonne on a 7E basis. Two figures from separate work, 44 years apart, landing in the same place. Table 01 carries three limits. McLaren and de Villiers were counting platinum, palladium, rhodium, ruthenium, iridium and gold, which is close to a 6E basis rather than a 7E basis. Their central eastern number rests on three cores. It is a regional average from 1982 rather than a measurement of this property, and it is independent rather than precise. Platinum and palladium are close to even here The same study reports the ratio of platinum to palladium by sector. The central eastern Bushveld runs at 1.21. The western limb runs at 2.45. Southern Palladium's Ore Reserve carries platinum at 2.34 grams per tonne and palladium at 2.33. A ratio of about 1.00. So the relative palladium richness that gives the company its name is a property of this part of the Bushveld rather than something particular to this project. On the western limb, platinum runs at roughly two and a half times palladium. On the eastern limb it runs close to even. The exposure a shareholder takes on the eastern limb is therefore differently weighted from the western limb. And there is more rhodium than in the Merensky Reef McLaren and de Villiers report rhodium at 9 per cent of the platinum group element and gold total in the UG-2, against 2 to 5 per cent for the Merensky Reef. Rhodium is the smallest number in the basket and often the largest in the revenue. Evolution Capital's model has rhodium at 22 per cent of net smelter return, level with palladium and behind only platinum.[R13] The long-term price assumptions in Southern Palladium's Optimised Pre-Feasibility Study of 10 July 2025[R2] put rhodium at US$6,190 an ounce against US$1,200 for platinum and US$1,100 for palladium. Those figures are the study's own long-term real assumptions, sourced by Minxcon to Consensus Economics, and they are stated here as the study's inputs rather than as current market prices. Southern Palladium's Ore Reserve carries rhodium at 0.48 grams per tonne within a 6E grade of 6.17, about 7.8 per cent. Close to what the 1982 work found across the whole reef. Grade is king, but not for the reason most people assume In the world of mining, our favourite phrase is that grade is king. Some people will equate a high grade to a sense of value. In reality, the grade doesn't really make the rock richer. What it does is allow you to move less dirt to get that value. That really boils down to an operating cost issue, and potentially a capital cost one as well. Hence, in my opinion, this is how it gives value to the project. Operating cost, usually shortened to opex, is what it costs to run a mine day to day. Capital cost, or capex, is what it costs to build it in the first place. The Optimised Pre-Feasibility Study proposes 1.2 million tonnes a year in Stage 1, rising to 2.4 million tonnes in Stage 2. That produces over 200,000 ounces of platinum group metals a year, rising to over 400,000. Chalice's Gonneville deposit carries 0.79 grams per tonne of 3E. To produce about 150,000 ounces a year in its Stage 1, Chalice's Pre-Feasibility Study proposes processing 1 million tonnes of oxide and 4 million tonnes of sulphide, rising to 1 million tonnes of oxide and 12 million tonnes of sulphide in Stage 2, and then to 14 million tonnes of sulphide from year 9. Southern Palladium moves 1.2 million tonnes for 200,000 ounces. Chalice moves 5 million tonnes for 150,000 ounces. Southern Palladium moves about a quarter of the rock and produces a third more metal. Section 8.00 puts a capital cost on that difference. 6.00 — THE E'S, AND WHICH ONES MAKE THE WORLD GO ROUND A twelfth of the metal, a third of the money Section 1.00 set out what the letters count. This section answers the second half of that question, which is which of those metals is actually earning the money. 3E counts platinum, palladium and gold. 4E adds rhodium. 6E adds ruthenium and iridium. 7E adds osmium. A company chooses which basis to report on, and the choice is not neutral. Reporting on 7E gives a bigger number than reporting on 3E for exactly the same rock. Southern Palladium's UG2 resource is 8.12 grams per tonne on 4E and 9.78 grams per tonne on 7E. Chalice reports Gonneville on 3E. Terra Metals reports the Dante Reefs on 3E. So the first thing to do with any platinum group element company is check which letters it is using before comparing its grade with anybody else's. The proportions, which is where it gets interesting The mix of metals in a deposit is called the prill split. It is simply what share of the total each metal makes up. Southern Palladium's Optimised Pre-Feasibility Study[R2] sets out what the mine is expected to produce across both stages over its 33-year life. Those figures, and the long-term prices the same study uses, are all that is needed to work out where the money comes from. Table 02 does that arithmetic. Table 02 divides out to a basket price of US$1,557 for every 6E ounce, which is the figure the study itself states. The arithmetic reconciles, so the inputs are the right ones. Rhodium is the one that makes the world go round Rhodium is 7.8 per cent of the ounces and 31.1 per cent of the revenue. It is the single largest earner in the basket, ahead of both platinum and palladium, on less than a twelfth of the metal. The reason is the price. The study assumes US$6,190 an ounce for rhodium against US$1,200 for platinum and US$1,100 for palladium. Rhodium is a very small market and a concentrated one. Johnson Matthey[R16], one of the world's largest refiners of these metals, reports that South Africa supplies over 80 per cent of the world's primary rhodium, ruthenium and iridium, and around 70 per cent of its platinum. The scale of the market is such that Johnson Matthey measures its annual balance in tens of thousands of ounces, recording a shortfall of 50,000 ounces in 2025 and forecasting a surplus of 15,000 ounces in 2026. Johnson Matthey groups platinum, palladium and rhodium together as the three autocatalyst metals, and forecasts rhodium demand falling about 6 per cent in 2026 as petrol car output declines. So rhodium is small, it comes overwhelmingly from one country, and what it is mostly used for is shrinking. The concentration runs in two directions. The upside is in section 5.00. McLaren and de Villiers found rhodium at 9 per cent of the UG-2 basket against 2 to 5 per cent for the Merensky Reef.[R7] The UG2 is the rhodium-rich reef, and Bengwenyama is a UG2 project. The risk is that a third of the revenue depends on the most volatile price in the group. Bridge Street's own assumed rhodium price has moved from US$6,000 an ounce in October 2024 to US$8,300 in July 2026 across four reports in under two years. A price that moves that far that fast can move back. Ruthenium is the one that flatters the ounce count Ruthenium is 12.7 per cent of the ounces and 3.7 per cent of the revenue. It is the second most abundant metal in the basket by weight and the second smallest by earnings. That is not a criticism of the company, which reports everything it has and states its basis clearly. It is a caution about the notation. Moving from a 3E number to a 6E number adds ruthenium and iridium. At Bengwenyama, 3E is 76.8 per cent of the ounces and 57.4 per cent of the revenue. So the extra metals counted in the 6E figure add about 23 per cent to the ounce count and about 13 per cent to the money. The 6E number counts metal that earns less per ounce than the metal in the 3E number. A 6E grade and a 3E grade are therefore not comparable figures. And you do not get the full price for any of it There is one more step between the metal in the ground and the money in the bank. A mine sells concentrate, and the smelter pays for a proportion of the metal in it rather than all of it. That proportion is the payability. Southern Palladium's Pre-Feasibility Study[R1] assumes 86 per cent for platinum and palladium. Bridge Street, working from their own market enquiries, use 80 per cent.[R12] That six percentage point difference applies to the two largest metals by weight in the basket, so it moves any project economics number that uses it. Payability for the minor platinum group metals is generally lower again, and in some contracts they are not paid for at all. So what is a good enough mix? The prill split on its own does not answer that. There is no combination of metals that is good in the abstract. What decides it is whether the basket those metals produce is comfortably more than what it costs to dig them up. The basket price is the way to boil the whole mix down to one number. It is the revenue from every metal, divided by the ounces, expressed as dollars per ounce. Bengwenyama's is US$1,557 for every 6E ounce on the study's assumptions, and that single figure carries the entire prill split inside it. Then you need something to measure it against. The Optimised Pre-Feasibility Study gives the breakeven basket price, which is the level the basket has to reach before the project makes nothing. Those figures are the most useful numbers in the whole study and neither research house quotes them. Table 03 sets them out. On the full two-stage project, and counting the cost of building it, the basket can fall 43 per cent from US$1,557 to US$891 before the project stops making money. That is what a good enough mix looks like. Not a particular arrangement of platinum, palladium and rhodium, but a basket with room to fall a long way and still work. The same test applies to any project, whatever letters it reports on. A low-grade deposit with a valuable mix and a high-grade deposit with a cheap one can arrive at the same place. And the mix still matters, even once you have the basket Two projects can have the same basket price and different risks inside it, so the split carries information the summary figure does not. Bengwenyama takes about a third of its revenue from rhodium. Rhodium and palladium both go overwhelmingly into catalytic converters for petrol engines. Platinum has a wider set of uses, including diesel catalysts, jewellery, industrial demand, investment, and a growing role in fuel cells and electrolysers.[R16] So more than half of the revenue here is exposed to one end market, and the most concentrated exposure of all is the rhodium third. Every UG2 mine on the Bushveld carries the same shape, and Bengwenyama's rhodium content is higher than the Merensky Reef's, which is an advantage while rhodium holds up. The rhodium price is the single largest variable in the basket. Support Samso → 7.00 — THE SECOND TEST, AN OREBODY THAT CAN BE MINED Seventy-three centimetres, and everything that interrupts it Bengwenyama has some real difficulties with the orebody itself. This section sets them out. The reef is narrow The UG2 at Bengwenyama averages about 73 centimetres. Three independent sources agree that a narrow reef is normal here, and Table 04 puts them alongside the company's own figure. A narrow reef means conventional narrow-reef mining, which is labour-intensive and has been practised on the Bushveld for a century. Southern Palladium plans a 1 metre stoping width, which means mining about 27 centimetres of waste with every 73 centimetres of reef. That dilution is already in the reserve grade. Plate 03 shows the seam and the cut the mine plan proposes to take. A dome has removed the reef from part of the property Roger Scoon and B. Teigler named the Eerste Geluk dome in 1995.[R9] It is a structure in the floor rocks beneath the Bushveld Complex that pushed upwards during and after the intrusion, disturbing the layers above it. The effect runs both ways. Where the reef survived, the dome lifted it closer to surface, which is part of why Bengwenyama is shallow. Where it did not survive, the reef is simply gone. Plate 04 is the company's own cross-section through the dome, and it carries the annotation “More than 600 m of Stratigraphy is missing (No MR, UG2, UG1 & LGs)”. Southern Palladium's drilling found the dome larger than the earlier modelling had assumed, and the affected ground was removed from the resource. There is a published precedent for what a dome can do Scoon returned to this subject in 2002, in a study of dome flanks in the northeastern Bushveld.[R10] He reported that uplift and attenuation on the flanks of domes are more extensive and more severe than previously realised, and that marker layers are absent from some drill holes with some missing altogether. At a property called Dwarsrand, an exploration programme found economically significant grades in Merensky Reef on a dome flank. Exploration was terminated because of poor strike continuity. Good grades, and the programme stopped for a reason that had nothing to do with grade. That is a documented outcome on this style of ground. Two accounts of the faulting, and they do not agree Southern Palladium reports a system of faults and dyke swarms with throws in excess of 200 metres, dividing the UG2 into 26 structural blocks.[R1] Scoon and Teigler reported something different in 1995. They found the Steelpoort fault, long thought to be the most prominent structural dislocation in the area, to be a comparatively minor feature. They found no evidence for its northward continuation into Eerste Geluk, and a vibroseismic survey on the property demonstrated a lack of major faulting. Faults cutting their drill-proven area had throws of less than 200 metres. These are different areas of the property and thirty years apart, with different survey coverage, so they are not necessarily in conflict. Both are reported here and neither is resolved. Potholes take about a fifth of the reef A pothole is a roughly circular depression in the reef where the layers have slumped or been scoured away. Inside one, the reef may be thinned, deformed or missing. Plate 05 shows the geometry. Plate 05 is a Merensky Reef example rather than a UG2 one, and the same geometry occurs on both reefs. Bengwenyama's UG2 is 77 per cent massive facies and 17 per cent pothole facies. Watson and colleagues reported in 2021 that potholes typically occupy 2 to 35 per cent of reef area for shaft blocks across the Bushveld.[R11] Bengwenyama's 17 per cent is within that range. The Pre-Feasibility Study applies geological losses of 21 per cent to Measured and Indicated material and 26 per cent to Inferred, with potholes accounting for roughly three quarters of those losses. Plate 06 maps where those losses fall across the property. What a loss factor does not capture is what potholes do to mining. Watson and colleagues report that potholes are generally left as stability pillars rather than mined, that stoping often stops short of the pothole edge, and that ground conditions deteriorate there with increased jointing. They measured rock strength inside a pothole against the same rock types outside one at the same mine. Norite came in at 50 megapascals inside against 118 outside. Pyroxenite at 45 against 100. Roughly half strength. They also measured stresses inside a pothole at 250 metres depth and found a major principal stress of 107 megapascals, against about 14 megapascals measured in the same rock type outside a pothole at similar depth. Plate 07 is what those numbers look like underground. The authors note this intensity of fracturing is not normally seen at 250 metres on the Merensky Reef, because the rock is usually strong enough at that depth to resist it. Two things about that measurement. The authors state it is the only such measurement they are aware of in a Bushveld pothole, so it is one documented case rather than a general finding. And it was taken on the Merensky Reef rather than the UG2. The depth of 250 metres is within Bengwenyama's early mining range. Their conclusion is that none of the current theories of how potholes form adequately explains the stresses they measured, and those theories cannot be used to predict how potholes will behave at depth. Potholes are mined around routinely across the Bushveld, and after decades of work the profession does not agree on how they form. Depth, and one decision that has been revised The initial mining area runs from surface to about 250 metres. For a Bushveld PGM project that is shallow, and it is the single largest advantage Bengwenyama has over deeper Eastern Limb operations. Plate 08 shows the depth profile across the reef. The Pre-Feasibility Study planned to intersect the reef at about 50 metres using a decline at 6 degrees. Later drilling moved that to about 80 to 85 metres, and the decline gradient to 9 degrees. The company's explanation is that a steeper decline reaches 80 metres for only marginally more cost than a shallower one reaches 50 metres, and that the deeper intersection avoids oxidised ore. Something turned up in the literature that the company does not mention, and it may work in its favour. Gain reported in 1985 that the ratio of platinum to platinum plus palladium in the UG-2 ran at 1.77 in the surface environment against 0.99 unoxidised.[R8] He cited published findings that palladium is more mobile than platinum under weathering conditions, so oxidation strips palladium out preferentially and leaves the remaining material relatively richer in platinum. If that holds at Bengwenyama, then avoiding the oxidised zone is not only about recovery. It would also mean the ore reaching the plant carries the metal split the resource model assumes, rather than a weathered version of it. Samso raises that as an observation, not a conclusion. Gain measured it at Maandagshoek forty years ago and no equivalent measurement at Bengwenyama has been published. Whether it applies here, and whether it is worth anything if it does, are questions for the company rather than answers Samso can give. It is put here because it was found in the published record and it looks like it could count in the project's favour. 8.00 — THE THIRD TEST, WHAT IT COSTS TO BUILD What it costs to buy an ounce a year The clearest way to compare two mining projects in different countries is to work out what each one costs to build for every ounce it will produce in a year. That figure removes the effect of scale. A big project and a small one can be put side by side and the question becomes what it costs to buy an ounce a year of production capacity. Both companies publish the numbers needed to work it out, so what follows is Samso's arithmetic on their figures rather than anyone's estimate. What Southern Palladium says it will cost The Optimised Pre-Feasibility Study of 10 July 2025[R2] puts Stage 1 initial capital at US$219 million. That builds a 1.2 million tonne a year operation producing over 200,000 ounces of platinum group metals a year. The capital breaks down as US$87 million for the plant, US$39 million for shared infrastructure, US$30 million for the tailings storage facility, US$27 million for direct mining, US$4 million for capitalised development and US$31 million of contingency. Stage 2 adds a further US$278 million to take throughput to 2.4 million tonnes a year and production to over 400,000 ounces, and the company intends to fund that out of Stage 1 cash flow rather than by raising it. Table 05 puts both projects side by side. What that works out to per ounce On Southern Palladium's figures, Stage 1 costs roughly US$1,100 for every ounce of annual production capacity. Across both stages the number is roughly US$1,300. On Chalice's figures, Stage 1 costs roughly A$5,500 an ounce, which is about US$3,900 at the exchange rate of 0.713 on 17 August 2026. Across the full project it is roughly A$7,000, or about US$5,000. So Bengwenyama is roughly three and a half times cheaper to build for each ounce a year it will produce. That difference comes from section 5.00. Southern Palladium is treating 1.2 million tonnes to get 200,000 ounces. Chalice is treating 5 million tonnes to get 150,000. A smaller plant costs less, and the grade is what allows the smaller plant. Two qualifications apply to that comparison. The ounces are counted differently. Southern Palladium's are 6E ounces, which count six metals. Chalice's are 3E, which count three. On a like-for-like basis the gap narrows, though it does not close. And Gonneville produces a great deal more than platinum group metals. Over its modelled life it also produces 160,000 tonnes of nickel, 186,000 tonnes of copper and 15,000 tonnes of cobalt. Loading all of Chalice's capital onto the platinum group ounces alone overstates the cost of those ounces, because a large part of that plant is there to recover the base metals. And here the comparison turns around Chalice's all-in sustaining cost is US$370 an ounce of 3E[R5], and US$50 an ounce over the first three years, both net of by-product credits. Southern Palladium's all-in sustaining cost is US$969 an ounce of 6E in Stage 1 and US$821 across both stages. Chalice describes Gonneville as the lowest cost platinum group metal mine in the western world and the lowest cost of any undeveloped project. Those two statements can both be true because they measure different things. Nickel, copper and cobalt revenue is subtracted from Chalice's costs before the cost per platinum group ounce is worked out. That is standard practice for a polymetallic mine and it is not a trick. It does mean Chalice's cost depends on nickel, copper and cobalt prices in a way Southern Palladium's does not. So one project is much cheaper to build and the other is much cheaper to run. Bengwenyama needs less money to get started. Gonneville makes more margin per ounce once it is going, provided its by-product prices hold. Neither of those is the whole answer, and an investor deciding between them is really deciding which risk they would rather carry. A funding risk at the front, or a base metal price risk for twenty-three years. Why the build cost is the one that decides things for a small company Southern Palladium's share price has ranged from A$0.195 to A$2.71 since May 2025, which on the shares now on issue is a market capitalisation somewhere between about A$26 million and about A$359 million. It closed at A$1.93 on 17 August 2026. Chalice reported A$63 million in cash and listed investments with no debt, and says it has enough funding to reach a final investment decision in 2028. A company of Southern Palladium's size raising the money for a US$219 million first stage is a large task. The same company raising A$820 million would be a different proposition entirely, and the eighth test in section 4.00 is about exactly that. Capital intensity is a property of the orebody and it does not change. Whether the money can be raised is a property of the market on the day, and Part 2 takes that up as the eighth test. 9.00 — WHAT WE KNOW SO FAR, AND WHAT IS LEFT Three tests answered, five still open Part 1 has run three of the eight tests. The three that have been done and the five that have not are different kinds of question. The three tests the rock has passed On grade, Bengwenyama measures up well. The company reports 9.78 grams per tonne on a 7E basis. A 1982 study by Mintek, sampling the UG-2 across the whole Bushveld and having nothing to do with this company, found the central eastern sector averaging 10.05 parts per million against a Bushveld-wide UG-2 average of 6.93. Bengwenyama is in that sector. Three separate measurements of reef width, from 1982, 1985 and 1995, agree with the company that the reef here is narrow, around 60 to 73 centimetres. On the basket, the mix works. Rhodium earns about a third of the revenue from less than a twelfth of the metal, which is a property of the UG2 and an advantage while rhodium holds its price. More usefully, the study's basket price of US$1,557 for every 6E ounce is 1.75 times the breakeven of US$891 across both stages. The basket can fall 43 per cent before the project stops making money. On what it costs to build, the numbers are good. Roughly US$1,100 of capital for every ounce of annual production capacity, against roughly US$3,900 for Chalice's Gonneville on the same calculation. The grade is what buys that. Southern Palladium treats 1.2 million tonnes to produce 200,000 ounces where Chalice treats 5 million to produce 150,000. What the same three tests also showed The orebody is difficult, and the difficulties are in the rock rather than in anything the company has done. What Part 1 has not touched Everything above is about the rock. None of it says whether this becomes a mine. Five tests remain. Whether the flowsheet is conventional or novel. Whether anyone will buy the concentrate. Whether the resource has been converted into a reserve that stands up. Whether the permits are in hand. And whether a company of this size can raise US$279 million. On four of those five, Southern Palladium has a good answer and the evidence is public. On one of them the answer is unresolved, and it is the one that stops more projects than any of the others. Part 2 takes all five, and gives the verdict. See what is inside → The vocabulary, in plain English References and sources Each source below carries an identifier, [R1] to [R16]. The same identifier appears in the body text at the point that source is used, so any claim can be traced back to the document it came from. Every figure in this piece labelled PLATE is reproduced from a published source and its caption credits that source. There are no original Samso illustrations in Part 1. All tables are compiled by Samso from the sources named in each table caption. Where a calculation is Samso's own, the table caption says so and the underlying figures are the companies' own published numbers. [R1] Minxcon (Pty) Ltd, “Southern Palladium Limited Pre-Feasibility Study, Summary Report”, effective date 23 October 2024. Source of Plates 01, 02, 03, 04, 06 and 08, and of the resource, reserve, geological loss and mine plan figures. [R2] Southern Palladium Limited, “Optimised Prefeasibility Study, Project NPV of US$857m with Improved Fundability Through Staged Development”, ASX and JSE announcement, 10 July 2025. Source of the staged development figures, capital costs, all-in sustaining costs, breakeven basket prices, life-of-mine production and the price deck. [R3] Southern Palladium Limited, “DFS metallurgical studies deliver outstanding PGM and Chromite recoveries”, ASX and JSE announcement, 9 July 2026. Source of the statement that no novel technology is being applied. [R4] Southern Palladium Limited, “Mining Right Granted for Southern Palladium's Bengwenyama World-class PGM-Chrome Project”, ASX and JSE announcement, 10 August 2026. [R5] Chalice Mining Limited, “Gonneville Palladium-Nickel-Copper Project Pre-Feasibility Study”, ASX announcement, 8 December 2025. Source of all Gonneville capital, production, cost, recovery and metallurgical test work figures. [R6] Chalice Mining Limited, Gonneville project page, chalicemining.com. Source of the Mineral Resource and Ore Reserve statements quoted. [R7] McLaren, C.H. and De Villiers, J.P.R. (1982), “The Platinum-Group Chemistry and Mineralogy of the UG-2 Chromitite Layer of the Bushveld Complex”, Economic Geology, volume 77, pages 1348 to 1366. [R8] Gain, S.B. (1985), “The Geologic Setting of the Platiniferous UG-2 Chromitite Layer on the Farm Maandagshoek, Eastern Bushveld Complex”, Economic Geology, volume 80, pages 925 to 943. [R9] Scoon, R.N. and Teigler, B. (1995), “A New LG-6 Chromite Reserve at Eerste Geluk in the Boundary Zone between the Central and Southern Sectors of the Eastern Bushveld Complex”, Economic Geology, volume 90, pages 969 to 982. [R10] Scoon, R.N. (2002), “A New Occurrence of Merensky Reef on the Flanks of the Zaaikloof Dome, Northeastern Bushveld Complex, Relationship between Diapirism and Magma Replenishment”, Economic Geology, volume 97, pages 1037 to 1049. [R11] Watson, B.P., Hoffmann, D. and Roberts, D.P. (2021), “Investigation of stress in potholes in the Bushveld Complex, a case study”, Journal of the Southern African Institute of Mining and Metallurgy, volume 121, number 1, pages 47 to 56. DOI 10.17159/2411-9717/1152/2021. Source of Plates 05 and 07. [R12] Bridge Street Capital Partners, research reports on Southern Palladium Limited, February 2024 to August 2026. Bridge Street are corporate advisers to Southern Palladium and lead manager on its initial public offering and capital raisings, and the analyst holds shares in the company. [R13] Evolution Capital, initiation of coverage on Southern Palladium Limited, 27 January 2026. Evolution Capital disclose that they have received a financial benefit from the company. [R14] Wagner, P.A. (1929), “The Platinum Deposits and Mines of Southern Africa”, Oliver and Boyd. [R15] Meeka Metals Limited (ASX: MEK), December 2025, March 2026 and June 2026 quarterly reports, and the Mt Holland South acquisition announcement. Production, cash flow and mine plan figures as reported in those quarterlies. The share price movement is from market data. Figures reached Samso through secondary coverage of the quarterlies and should be confirmed against the originals before publication. [R16] Johnson Matthey, PGM Market Report, May 2025 and May 2026 editions, matthey.com. Published annually and covering supply, demand and market balance for platinum, palladium, rhodium, ruthenium and iridium. Source of the South African supply shares, the rhodium market balance figures and the autocatalyst demand forecast. Cover photo: Bushveld Igneous Complex, South Africa (ASTER), NASA/METI/AIST/Japan Space Systems, and U.S./Japan ASTER Science Team. Public domain, via Wikimedia Commons (https://commons.wikimedia.org/wiki/File:Bushveld_Igneous_Complex,_South_Africa_(ASTER).jpg). Reproduced photograph, shown faded under a navy overlay on the article thumbnail. The Samso Way – Seek the Research Here at Samso, we pride ourselves on delivering content for investors that is independent and informed by over three decades of experience in the industry. Our content is well-researched and is only created if I see merit in discussing the company's story. Our mission is simple: cut through the noise and spotlight what matters—genuine stories, grounded insights, and real opportunity. Investors can explore our four core platforms: Coffee with Samso Samso Insights Samso News Samso Research There may be numerous paths to success in investing, but the common thread among successful individuals is that they remain committed to making informed decisions. Equip yourself with the right knowledge and tools, and you will be well on your way to achieving your financial goals. Most importantly, investors need to be absolutely diligent in understanding their own risk-reward tolerance and capabilities. Never bite off more than you can chew. As they say, Rome wasn't built in a day, and the Great Wall stood because it took centuries to complete. The Samso Philosophy: Stay curious. Stay sharp. And remember—digging deeper always uncovers the real value. In Life, there is no such thing as a Free Lunch. Never bite off more than you can chew is my parting comment. Happy Investing, and the only four-letter word you need to know is DYOR. To support our independent work, please head over to our Support Page and give us a helping hand in any of the ways listed. This is a new initiative for the Samso Platform, and it was always the concept of Samso when we started this journey in 2018. Disclaimer This article is general information only. It is not financial product advice and it does not take into account any reader's objectives, financial situation or needs. It contains no price targets, no valuations and no recommendation to buy, hold or sell any security. Companies named in this article are used as historical illustrations of events that have already occurred and no view is expressed on their current merits. Readers should consider obtaining advice from a licensed financial adviser and should conduct their own research before making any investment decision. Samso and its associates may hold interests in companies mentioned. Read full Disclaimer. About Samso Samso is a trusted platform that equips dedicated investors with up-to-date industry knowledge and insights from top CEOs and thought leaders. By staying informed on business advancements and market trends, investors can enhance their financial decisions through a combination of expert guidance and their own research. The Investor's Playbook is an open-ended, evergreen educational series about how to think about investing rather than what to buy. Samso Insights | www.samso.com.au | An Investor Lens on ASX-Listed Companies

AI's inputs got more than 99 per cent cheaper, and the bills kept rising. This is what agentic AI actually costs, in plain English, and the one test to apply to any AI story a company tells you. Samso Insights Research Insights Tokens and AI Samso Investing Series A note on the vocabulary and the sources. Technical and mining terms are explained in plain English where they first appear, and all of them are collected in one box at the back. Every source carries an identifier, [R1] to [R3], and that identifier appears in the body text at the point the source is used, so any figure can be traced to where it came from. WHAT THIS PIECE COVERS 1.00 Cheaper by the unit, dearer by the invoice What a token is, and the puzzle of a falling price beside a rising bill. 2.00 What a Quadrillion Tokens Looks Like The scale of demand, drawn from what Google reported at I/O 2026. 3.00 Where the Money Actually Goes The six reasons a cheaper token still produces a bigger invoice. 4.00 Tokens Are the Bill, Not the Value The distinction the whole note rests on, and why cost per outcome is the number. 5.00 Six Questions Before You Believe an AI Story The framework turned into questions an investor can put to a company. 1.00 — THE PARADOX Cheaper by the unit, dearer by the invoice Start with tokens. Everything in this note runs on them. A token is a small chunk of text, roughly a short word or part of a word. It is the basic unit an AI model reads and writes. Tokens count for three reasons. With that in hand, here is the puzzle that frames everything else. The price of that raw input has collapsed, and yet company spending on AI has gone up, not down. On the price, the direction is not in doubt. Stanford University's 2025 AI Index reported that the cost of running a model at the level of GPT-3.5 fell from about US$20 per million tokens to about US$0.07 through 2024, a fall of more than 99 per cent. Over the same broad period, the venture capital firm Menlo Ventures estimated that enterprise spending on large language models (the kind of AI behind chat assistants and coding tools) tripled in the twelve months to the end of 2025. The price per unit fell through the floor. The total bill still multiplied. FIG. 01 sets the two lines side by side. [R1] The instinct is to call this waste. That instinct is worth resisting. A falling unit price alongside rising total spending is the normal signature of a technology being adopted quickly, not proof that money is being burned. Cars became cheaper to run per kilometre over the twentieth century, and total spending on motoring rose all the same because people drove far more. The real question is not whether AI spending is rising. It is whether each dollar of it is buying something worth having. Holding that question in mind is the whole discipline this note is about. 2.00 — THE SUPPLY SIDE What a Quadrillion Tokens Looks Like The scale the bill is paying for If the first section is the demand side, the clearest picture of the supply side came from Google's I/O 2026 developer conference. The numbers are difficult to hold in your head. Google reported that the volume of tokens processed across its products went from 9.7 trillion a month two years ago, to about 480 trillion a month a year ago, to about 3.2 quadrillion a month now. A quadrillion is a thousand trillion, so the latest figure is roughly seven times the level of a year earlier. FIG. 02 shows the shape of that climb. The line barely lifts off the floor for most of the period, then goes near vertical, which is the visual an investor should associate with the current phase of AI adoption. [R2] The demand behind that curve is not only consumer. A companion figure Google showed, redrawn here as FIG. 03, tracks tokens processed through its model APIs, the pipes that developers and businesses build on. That volume rose from about 7 billion tokens per minute in September 2025 to about 19 billion per minute by April 2026, roughly six times higher year on year. This is the line to watch most closely for investors, because it is the closest proxy for businesses actually wiring AI into their operations, which is where the enterprise bill from Section 1 comes from. That demand rides on enormous distribution. Google said it now runs 13 products with more than a billion users each, five of them with more than three billion users (Search, Gmail, Android, Chrome, and YouTube). Its Gemini AI app alone roughly doubled its monthly users in a year, from about 400 million to more than 900 million. To serve it, the company's capital expenditure, the money spent building data centres and buying chips, is guided to rise about six-fold in four years, as FIG. 04 shows. That capital line is the bill from Section 1, seen from the side of the company sending the invoices. The tokens counted so far are produced inside physical facilities like the one in PLATE 1, which is a powered, cooled hall filled with computers. [R3] SAMSO ACCESS Some of the work does not appear here Samso holds back a set of Insights for members. Same method as this piece, the same sourcing and the same counterweights, on companies and sectors that do not make the public feed. Every member-only Samso Insight unlocked Full access to the gated Insights archive New member-only pieces included as they publish Samso News and Research Workups will join the gated set, with Research Workups also available on their own tier. A$20 a month · cancel any time See what is inside → 3.00 — WHERE THE MONEY GOES Where the Money Actually Goes Why a cheap token still adds up to a fortune If tokens are so cheap, how does the bill get so large? McKinsey's answer is that the cost of AI has moved out of the price per token and into the way the work is done, especially once the software starts acting on its own. [R1] The relevant shift is from a model that answers a single question to an AI agent, which is software that chains many AI steps together, uses tools, and checks its own work to finish a task without a human at each step. That autonomy is where the money goes. McKinsey set out several reasons, summarised in FIG. 05. In plain terms, an agent re-sends its full context to the model again and again as it works. Much of what it does is checking and re-doing its own output. Its freedom to choose a path means the same job can cost very different amounts on different runs. Powerful, expensive models get used on trivial steps. Coordinating several agents multiplies the traffic. And messy data or long-winded prompts inflate the count before any useful work begins. Google's own keynote gave the perfect worked example, shown in FIG. 06. Its engineers set an agent system loose to build a working computer operating system from an empty project. Over 12 hours, 93 sub-agents made more than 15,000 model requests and processed 2.6 billion tokens, and produced the core of a functioning operating system for under US$1,000 of usage. Read that two ways at once. It is astonishingly cheap for work that would take skilled people many months. It is also 2.6 billion tokens for a single task, and multiplied across millions of tasks that is exactly how you get to a US$180 billion capital bill. [R2] COFFEE WITH SAMSO The same companies, on the record Coffee with Samso is the interview series. Management answering questions in long form, so the people behind a project can be judged alongside the numbers in a piece like this one. Free to watch. Companies commission an interview, and it stays free to everyone, permanently. Watch the interviews → 4.00 — THE BILL AND THE VALUE Tokens Are the Bill, Not the Value The one line to carry through every AI announcement Here is the idea the title promises, stated plainly. Tokens, models, and dollars spent are inputs. They are the bill. Whether any of it produced something worth having is a separate question, and it is the only one that counts. McKinsey's phrase for the mistake is judging AI by its inputs instead of its outcomes. The number they argue companies should track is cost per outcome, which is the cost of a finished, useful result, a resolved customer complaint, a cleared invoice, a piece of code shipped, rather than the cost of the tokens burned along the way. FIG. 07 draws the distinction. Most organisations cannot yet produce that number. In a McKinsey survey on AI financial operations from May 2026, 93 per cent of respondents said they had exceeded their AI budgets, and in a separate McKinsey survey about one in five said their organisation had limited its use of AI because of running costs. [R1] Those figures sound alarming, and they need their counterweight. Budgets set at the dawn of a new technology are guesses, so overshooting them is not by itself evidence of failure. It is evidence that very few organisations yet measure the thing that would tell them whether the spend was worth it. That gap, between spending confidently and measuring plainly, is the gap an investor can look through. SUPPORT SAMSO Independent, organic, and free to read We produce independent organic content and need your support to keep that going. Every contribution covers the cost of research, keeps the site running, and buys the time to write the next piece. A$10, A$20, A$30 or an amount you choose One time, weekly, monthly or yearly A donation buys nothing and unlocks nothing. It keeps the work going. Support Samso → 5.00 — SIX QUESTIONS Six Questions Before You Believe an AI Story Turning the framework into investor diligence McKinsey wrote its recommendations for chief executives. Turned around, they become a short checklist for anyone deciding whether a company's AI story deserves to move the share price. None of these need technical knowledge to ask. [R1] What did the spend buy? Ask for the outcome in cost-per-outcome terms, not tokens, licenses, or headcount "freed up". A company that can answer this is rare and worth noting. Where does AI actually change the economics? A serious operator can name the few parts of its business where machine work alters the cost or quality in a way that counts, rather than claiming AI everywhere. Who owns it, and what are they measured on? Look for clear accountability for AI spending and results, with real targets, not a vague "innovation" mandate. Is the spend concentrated or sprayed? Heavy AI use tends to cluster in a few high-value places. At McKinsey's own firm, about 10 per cent of users drove roughly 65 per cent of token use. Concentration can be healthy focus or hidden dependence; the point is that management should know which. Is the tool matched to the task? Using the most powerful, most expensive model for everything is a red flag. McKinsey found that simply encouraging shorter prompts and answers cut token use by 30 to 40 per cent in some workflows without materially hurting quality. Can they separate the bill from the value? If management cannot tell you, in plain numbers, what the AI cost and what it returned, you are being asked to take the value on faith. One caution keeps these questions fair. Most companies cannot yet answer most of them, so a string of "no" answers is common today and not automatically damning. The tools to measure this are themselves new. The signal to pay for is a company that answers them well. There is also a subtler point from McKinsey. As AI makes clever workflows easy to copy, a slick "we use AI" process is a weaker moat than it looks. Proprietary data a rival cannot copy and the discipline to run AI cheaply become the real, and quieter, advantages. A company that never headlines AI may hold more of it than one that never stops talking about it. 6.00 — REFERENCES AND SOURCES Where every figure in this note came from This note draws on one primary article and one primary event transcript, listed below. The AI Index, Menlo Ventures, and arXiv figures reach us as reported within the McKinsey article. Where a reader wants them stood behind their original sources, those originals should be consulted directly. All figures shown in the visuals are original Samso illustrations of the data named in each caption. Market-sensitive and time-stamped figures (token volumes, survey dates, capital-expenditure guidance) are quoted as at the dates shown and should be refreshed on publication day. Photographs labelled PLATE are reproduced from a company's public ASX release, with attribution in the caption, and are distinct from the original Samso illustrations labelled FIG. [R1] McKinsey & Company (QuantumBlack) — "Is that AI agent worth it? Agentic economics and the modern operating model" (2026). Source of: the token price fall (US$20 to US$0.07 per million tokens through 2024, attributed to the Stanford HAI 2025 AI Index); enterprise LLM spending tripling to end 2025 (attributed to Menlo Ventures); the six cost drivers; the ~1,000× token multiple, ~60 per cent refinement share, and factor-of-30 variation (attributed to arXiv research); the "cost per outcome" concept; 93 per cent exceeding AI budgets (McKinsey Enterprise AI FinOps survey, May 2026); about one-fifth constraining AI use on cost (McKinsey State of AI survey); the roughly 10 per cent of users driving about 65 per cent of consumption; and the 30 to 40 per cent saving from more concise prompts. [R2] Google I/O 2026 keynote — transcript supplied by Samso, with the "Monthly Tokens Processed" and "Tokens Per Minute Processed" slides. Source of: 9.7 trillion, ~480 trillion, and ~3.2 quadrillion monthly tokens; ~19 billion tokens per minute across model APIs; 13 products with 1 billion-plus users and five with 3 billion-plus; the Gemini app rising from ~400 million to more than 900 million monthly users; capital expenditure of ~US$31 billion (2022) rising to ~US$180 billion to US$190 billion (2026); and the agent demonstration (93 sub-agents, 15,000-plus model requests, 2.6 billion tokens, under US$1,000, 12 hours). The 3.2 quadrillion monthly-token figure is corroborated by third-party coverage of I/O 2026 (e.g. Crypto Briefing; Thurrott; Google's own I/O 2026 blog posts). [R3] DigiCo Infrastructure REIT (ASX: DGT) — 1H FY26 Results Presentation (20 February 2026): source of the data-hall photograph reproduced in PLATE 1, with attribution. Cover photo: Front of server racks at NERSC, Derrick Coetzee from Berkeley, CA, USA. CC0, via Wikimedia Commons (https://commons.wikimedia.org/wiki/File:Front_of_server_racks_at_NERSC.jpg). Reproduced photograph, shown faded under a navy overlay on the article thumbnail. The Samso Way – Seek the Research Here at Samso, we pride ourselves on delivering content for investors that is independent and informed by over three decades of experience in the industry. Our content is well-researched and is only created if I see merit in discussing the company's story. Our mission is simple: cut through the noise and spotlight what matters—genuine stories, grounded insights, and real opportunity. Investors can explore our four core platforms: Coffee with Samso Samso Insights Samso News Samso Research There may be numerous paths to success in investing, but the common thread among successful individuals is that they remain committed to making informed decisions. Equip yourself with the right knowledge and tools, and you will be well on your way to achieving your financial goals. Most importantly, investors need to be absolutely diligent in understanding their own risk-reward tolerance and capabilities. Never bite off more than you can chew. As they say, Rome wasn't built in a day, and the Great Wall stood because it took centuries to complete. The Samso Philosophy: Stay curious. Stay sharp. And remember—digging deeper always uncovers the real value. In Life, there is no such thing as a Free Lunch. Never bite off more than you can chew is my parting comment. Happy Investing, and the only four-letter word you need to know is DYOR. To support our independent work, please head over to our Support Page and give us a helping hand in any of the ways listed. This is a new initiative for the Samso Platform, and it was always the concept of Samso when we started this journey in 2018. Disclaimer This article is general information only. It is not financial product advice and it does not take into account any reader's objectives, financial situation or needs. It contains no price targets, no valuations and no recommendation to buy, hold or sell any security. Companies named in this article are used as historical illustrations of events that have already occurred and no view is expressed on their current merits. Readers should consider obtaining advice from a licensed financial adviser and should conduct their own research before making any investment decision. Samso and its associates may hold interests in companies mentioned. Read full Disclaimer. About Samso Samso is a trusted platform that equips dedicated investors with up-to-date industry knowledge and insights from top CEOs and thought leaders. By staying informed on business advancements and market trends, investors can enhance their financial decisions through a combination of expert guidance and their own research. Samso Insights | www.samso.com.au | An Investor Lens on ASX-Listed Companies

All prices in this piece are as at the close of ASX trading on 8 October 2026. Global prices for gold are taken at the same time. Contents 6 The Geology of Mt Gibson 11 Capital, Operating Costs and the Study Economics 7 The Mt Gibson Mineral Resource 12 Funding, Approvals and the Timetable 8 The Mt Gibson Ore Reserve 21 References and Sources 9 The Mt Gibson Mine Plan 22 Disclaimer 10 Processing and Metallurgy 6 The Geology of Mt Gibson Mt Gibson is a long, steep, low-grade gold system in old volcanic rock. The gold sits in a set of parallel sheets that run north to south for about 8 kilometres. Drilling has followed them 950 metres down the slope of the sheets and they are still mineralised at that depth.[R2] Most of the tonnes are wide and low in grade, which suits a large open pit. Inside the sheets are narrower, richer shoots, and the underground plan is built on those.

An impressive high-grade 4.20 metre intersection at 42.43g/t Au — including 1.37 metres at 126.93g/t Au — has reinforced the case for Burbanks. A second drill rig is now pushing the Inferred Resource down to 450 metres below surface, and a paired tenement swap with Accelerate Resources delivers fresh ground proximal to the Black Swan Processing Hub. Samso News | ASX: MRT | Gold | Black Swan Processing Hub | Burbanks Gold Project | Maritana Minerals Limited (ASX: MRT), formerly Horizon Minerals Ltd, has progressed three significant work streams across its Western Australian gold portfolio in May 2026. The activity sits within the company's broader strategy to refurbish the 2.2 Mtpa Black Swan Processing Hub and bring it back online as a standalone gold producer, with first gold targeted for mid-2027. Earlier this month, the company delivered additional high-grade assay results from its completed Phase 1 diamond drill program at the 100%-owned Burbanks Gold Project. Bolstered by these results, Maritana kicked off phase 2 drilling at Burbanks mid-May , with a second rig mobilised to site. On 19th May, the company announced a dual tenement transaction with Accelerate Resources Limited (ASX: AX8), acquiring the Kanowna East tenure package and divesting its non-core Balagundi tenure package. Table 1: Summary of May 2026 activities Samso Summary Details Company Maritana Minerals Limited ASX Code MRT Primary Assets Black Swan Processing Hub (2.2Mtpa, in C&M); 1.9Moz Group Mineral Resource across multiple WA Goldfields projects Main Story Phase 1 drilling delivered ultra-high-grade results at Burbanks; Phase 2 drilling now underway with second rig mobilised; Kanowna East acquired and Balagundi divested in dual transaction with Accelerate Resources Operating Platform Fully funded through A$175M placement plus A$4.65M SPP; A$114.1M cash and no debt (as at 6 March 2026) Black Swan Studies Pre-production capital A$160.5M; capital intensity ~A$1,500/oz; project pre-tax NPV8 ~A$631M; IRR 83% at A$5,500/oz gold price Burbanks Phase 1 53 holes for 15,798m completed; standout intercept 4.20m at 42.43g/t Au including 1.37m at 126.93g/t Au from 290.84m in 26HBBD044 Burbanks Phase 2 Designed to extend Inferred Mineral Resource from 250m to up to 450m below surface; revised total program 25,000m Kanowna East Acquired from Accelerate Resources for ~A$200,000 in MRT shares (net); proximal to Gordon's Dam and Black Swan Balagundi Divestment Divested to Accelerate; Maritana retains 1% NSR royalty and a 5-year right of first refusal over future toll treatment at Black Swan Burbanks Phase 1 Drilling: High-Grade Results Define the Story The Burbanks Gold Project is located on a granted Mining Lease (M15/161) nine kilometres south of Coolgardie in the Western Australian Goldfields. The mineralisation is hosted within the Burbanks Shear Zone, a five-kilometre strike package of basalts and intercalated gabbro, dolerite and sediments (Figure 1). Gold occurs in ptygmatically folded and boudinaged laminated quartz veins with pyrite, pyrrhotite, scheelite and an alteration assemblage of plagioclase, calcite, biotite and garnet. Historical underground mining at Burbanks produced 324,000 ounces at a head grade of 22.7g/t gold between 1885 and 1961. In geology, a ptygmatically folded rock layer or vein features highly irregular, tight, and complexly contorted wrinkles that resemble a chaotic, looping snake or a squeezed tube of toothpaste. (Click Here for illustration) A boudinaged rock layer or vein has been stretched and pulled apart into a series of sausage-shaped fragments (Click Here For Illustration) Figure 1: Burbanks Gold Project showing the location of the Phase 1 completed holes and the current Phase 2 drill program (Source: MRT ASX Announcement 14 May, 2026). During the phase 1 program, the company reported a series of high-grade results that have materially improved the geological understanding of the deposit. Hole 26HBBD044 is the standout. It returned a 4.20 metre intersection at 42.43g/t gold from 289.63 metres, which includes 1.37 metres at 126.93g/t gold from 290.84 metres. The same hole returned a 0.30 metre intercept at 54.02g/t gold from 207.41 metres and a 1.84 metre intercept at 5.98g/t gold from 223.00 metres, including 1.20 metres at 8.69g/t gold (Figure 2). Figure 2: Gold mineralisation located within a quartz vein hole 26HBBD044. The interval grades 4.20m @ 42.43g/t Au from 289.63m including 1.37m @ 126.93 g/t Au from 290.84m (Source: MRT ASX Announcement 04 May, 2026). Maritana began a two-phase extensional and infill drilling program at Burbanks in late June 2025. Phase 1 was designed to improve geological confidence at the Burbanks North open pit, which is included in the current LOM plan, and at the Burbanks underground, which is not. Phase 1 comprised 53 diamond drill holes for a total of 15,798 metres. Assay results have been received for all but the final four holes, with two further results pending at the time of the 4 May 2026 release. The company also highlighted that visible gold was logged in quartz veining within the shear zone in 26HBBD044, with a photograph of the mineralisation included in the announcement. The presence of visible gold in a hand-sized core sample is a strong geological indicator but introduces the well-known nugget effect into grade interpretation. From a project assessment perspective, the result is significant because the standout intercept is associated with the broader Burbanks Shear Zone rather than a single anomalous nugget. Other notable Phase 1 results include a 6.49 metre intersection at 2.66g/t gold from 354.51 metres in hole 26HBBD039 (including 1.27 metres at 6.54g/t gold), a 1.81 metre intersection at 13.82g/t gold from 113.88 metres in 26HBBD040, a 2.02 metre intersection at 5.99g/t gold from 244.98 metres in 26HBBD042, and a 0.63 metre intersection at 24.32g/t gold from 460.37 metres in 26HBBD038 including 0.30 metres at 40.95g/t gold. Table 2: Selected significant intercepts from Phase 1 drilling at Burbanks Hole ID From (m) Interval (m) Au (g/t) 26HBBD044 289.63 4.20 42.43 incl. 290.84 1.37 126.93 26HBBD044 207.41 0.30 54.02 26HBBD044 223.00 1.84 5.98 incl. 223.64 1.20 8.69 26HBBD038 460.37 0.63 24.32 incl. 460.37 0.30 40.95 26HBBD039 354.51 6.49 2.66 incl. 355.97 1.27 6.54 26HBBD040 113.88 1.81 13.82 26HBBD042 244.98 2.02 5.99 26HBBD049 229.62 0.52 10.83 Intercepts are downhole intervals and not true widths. Reporting uses a 1.0g/t Au cut-off with a maximum of 1m internal dilution. No top-cut has been applied. Full results are reported in the 4 May 2026 ASX announcement. Beyond the headline grades, the spatial context of the Phase 1 pierce points is the more material outcome. The schematic long section (Figure 2A) published in the 4 May 2026 release shows that the recently reported intercepts sit both within and outside the current Inferred Mineral Resource envelope. The deep intercept in 26HBBD038 (0.63 metres at 24.32g/t Au from 460.37 metres) is well below the existing resource shell, supporting the case that mineralisation continues at depth. This is the geological basis for the Phase 2 program now underway. Figure 2A: SW NE schematic long section on local grid showing the location of drilling in the 2025-26 drill program at Burbanks North, the current MRE categories, historical pierce points and recent drill traces with pierce points >1.0g/t. Previously reported intersections from this program are highlighted in white boxes, current intersections are in yellow. (source Maritana Minerals Limited - ASX Release 4th May 2026) Management Commentary — Phase 1 Results “Following strong results previously reported, the latest assays from Burbanks continues to demonstrate consistent high-grade mineralisation and improved geological confidence. The ongoing success of the drilling program continues to de-risk the project, as we progress Burbanks towards becoming a reliable source of feed for Maritana's 100% owned Black Swan Processing Hub.” — Grant Haywood, Managing Director and CEO, Maritana Minerals Limited (4 May 2026) Burbanks Phase 2: Drilling the Resource at Depth Maritana kicked off Phase 2 drilling in the days following the 4 May 2026 Phase 1 results release. A second drill rig has been mobilised to site to accelerate the program (Figure 3). The two-phase program has been revised to a total of 25,000 metres, comprising both reverse circulation (RC) and diamond drilling The Phase 2 program is designed to test and extend the Inferred Resource of the high-grade Burbanks Main and Birthday Gift Lodes from the current depth of 250 metres below surface to up to 450 metres below surface. The company noted that access to neighbouring tenements held by Auric Mining was required to support optimal drill positioning, given the steeply dipping geometry of the orebody and the limited width of Maritana's own tenure. This access has been secured, allowing drill collars to be stepped back to test the mineralisation at depth. Figure 3: - Phase 2 drilling commenced on Burbanks Main and Birthday Gift Lodes. Second drill rig mobilised MRE Underway All assay results from Burbanks North drilling have been received. The company has stated that an updated Mineral Resource Estimate is now underway and scheduled for completion in the June 2026 quarter. Ore Reserve studies based on the improved confidence in resource classification from Inferred to Indicated will then follow. From a project assessment perspective, the importance of Phase 2 is that Burbanks underground was not included in the Black Swan Processing Hub Scoping Study. The release describes this as significant upside potential to the current LOM plan, which is targeted to produce 100,000 ounces of gold per annum over an initial period of five years. The Phase 2 drilling is the work program designed to test whether the depth extension can carry the high grades reported from Phase 1. Management Commentary — Phase 2 Commencement “Maritana's Phase 1 drilling at the Burbanks Gold Project has reaffirmed the high-grade nature of the mineralised system and increased our geological confidence in the deposit. "Our Phase 2 program is the next step to accelerating our exploration and development strategy for the Project, improving our current mine plan to feed the Black Swan Processing Hub.” — Grant Haywood, Managing Director and CEO, Maritana Minerals Limited (14 May 2026) Burbanks South: A Second Target Area Opens Up Alongside the Phase 2 deep drilling, Maritana has initiated exploration activity at Burbanks South, an area targeting high-grade shallow gold mineralisation. The 14 May 2026 release noted that preliminary observations from recent drilling indicate the presence of shearing and quartz veining consistent with the structural controls observed at Burbanks North. The company reported that the area is supported by numerous historic workings that highlight its prospectivity, but that limited shallow drilling completed to date means the extent and continuity of mineralisation remains largely untested. Follow-up RC drilling is being planned to test mineralisation along strike, with geological modelling underway to refine targets. Burbanks South remains at the target-generation stage. The next test is whether follow-up RC drilling can confirm a structural and grade footprint that justifies elevating it to a defined target along the broader Burbanks corridor. Kanowna East: A Strategic Tenement Swap with Accelerate Resources On 19th May, Maritana Resources executed a dual tenement transaction with Accelerate Resources Limited (ASX: AX8). Under the agreements, Maritana will acquire Accelerate's Kanowna East tenure package and will divest its own Balagundi tenure package to Accelerate. The acquired Kanowna East package comprises twelve tenements, including five granted tenements (E27/596, E27/700, E27/704, P27/2428 and E27/752) in which Accelerate holds a 70% interest with Metal Hawk holding the remaining 30%, and seven tenements in application (E27/743, E27/750, E27/754, P27/2642, P27/2643, P27/2644 and P27/2645) in which Accelerate holds a 100% rights-in-application interest. The acquired ground is described by the company as proximal to its existing Kalgoorlie operations, including the Gordon's Dam Project and the Black Swan Processing Hub (Figure 4). Figure 4: Greater tenement area around Black Swan Operations, showing the acquired Kanowna East tenements (green), the existing Gordon's Dam Project, the Wilsons Prospect and the Black Swan Processing Hub. (Source: Maritana MRT ASX announcement 19 May, 2026) Net consideration under the transaction comprises A$200,000 payable in Maritana shares, calculated as the greater of 133,333 shares or shares worth A$200,000 based on Maritana's five-day VWAP before execution. The cash consideration payable under each sale agreement is A$60,000, with the two amounts effectively set off against each other. Completion of the transaction is expected in the June 2026 quarter, and the acquired tenements will be incorporated into Maritana's FY2027 exploration strategy. A 1% net smelter return royalty applies to all minerals extracted from each side of the transaction, granted in favour of the divesting party. This is a standard royalty structure for tenement transactions of this type and ensures both parties retain a financial interest in any future production from the ground they have transferred. Balagundi: Divestment Preserves Optionality on Future Toll Treatment The Balagundi tenure package divested to Accelerate comprises seventeen tenements held by Kalgoorlie Ore Treatment Company Pty Ltd, all of which are granted. The release described Balagundi as non-core to Maritana's current Kalgoorlie operating district. The divestment removes ongoing exploration expenditure commitments from a part of the portfolio that does not fit the company's current operational focus. Importantly, Maritana has retained a structural right over future toll treatment of Balagundi ore at the Black Swan Processing Hub. Under the Balagundi sale agreement, Accelerate has been granted a five-year right of first refusal in respect of future toll treatment opportunities at Black Swan, subject to the terms of the agreement. Maritana has also granted Accelerate the benefit of the existing Land Use Agreement with the Applicants for the Marlinyu Ghoorlie Native Title Claim. The practical effect is that if Accelerate progresses Balagundi toward development, Maritana has structural exposure to that ore as a potential future feed source for Black Swan. The company has framed this as preserving strategic optionality without continuing to carry the asset on its balance sheet. Management Commentary — Kanowna East and Balagundi “This transaction reflects Maritana's disciplined approach to portfolio management. By acquiring the Kanowna East Project and divesting the Balagundi tenure, we are consolidating our landholdings around the Black Swan Processing Hub, while rationalising assets that are better suited to a party with the scale and intent to advance them. "Importantly, the right of first refusal we have retained over future toll treatment at Black Swan preserves exposure to a potential future ore source and revenue stream open for Maritana shareholders, should Balagundi progress to development. We consider this a sensible, value-accretive outcome for both companies.” — Grant Haywood, Managing Director and CEO, Maritana Minerals Limited (19 May 2026) Black Swan Processing Hub: The Operating Anchor The Black Swan Processing Hub is the central asset around which Maritana's standalone gold producer strategy is built. The plant is located approximately 50 kilometres north-east of Kalgoorlie in Western Australia and has a nameplate capacity of 2.2Mtpa (Figure 5). It is currently in care and maintenance, having previously operated as a nickel sulphide flotation facility. Figure 5: Maritana's project locations, regional geology and surrounding infrastructure, including the Accelerate Resources tenement acquisitions. (Source: Maritana Minerals ASX announcement dated 19 May 2026) The Black Swan Scoping Study released by the company on 17 February 2026 supports a conversion of the existing comminution circuit to gold processing with the addition of a carbon-in-leach (CIL) circuit. The estimated replacement value of the existing infrastructure is more than A$200 million, against a stated capital cost of A$101 million for the plant refurbishment and approximately A$160.5 million in total pre-production capital. Capital intensity is approximately A$1,500 per ounce. Headline study outcomes include average production of 102,000 ounces per annum, life-of-mine production of 546,000 ounces over five years, C1 cash cost of A$2,852 per ounce, all-in sustaining cost of A$3,353 per ounce, and a project pre-tax NPV8 of approximately A$631 million at an A$5,500 per ounce base case gold price. The study reported an IRR of 83% and a payback period of 18 months from plant commissioning. Maritana is fully funded for the refurbishment and the commencement of mining and processing through an A$175 million placement completed in February 2026, a A$4.65 million Share Purchase Plan completed in March 2026, and existing cash and listed investments. The company reported A$114.1 million in cash and cash equivalents as at 6 March 2026, with no debt. Near-Term Milestones to Watch The following near-term milestones have been disclosed across the three ASX announcements and the April 2026 investor presentation: Table 3: Near-term milestones Activity Timing Status / Source Burbanks Phase 1 remaining assays (final 4 holes including 26HBBD052 and 26HBBD053) Underway Pending at time of release (4 May 2026 release) Burbanks North Mineral Resource Estimate update June 2026 quarter All Phase 1 assays received; MRE update to enable Inferred to Indicated reclassification (14 May 2026 release) Burbanks Phase 2 drilling — Main Lode and Birthday Gift Lodes from 250m to 450m below surface Underway Commenced; second rig mobilised (14 May 2026 release) Burbanks Ore Reserve studies Following MRE update To follow Inferred to Indicated reclassification (14 May 2026 release) Burbanks South RC drilling program Planned Geological modelling underway; follow-up RC drilling planned (14 May 2026 release) Kanowna East tenement acquisition completion June 2026 quarter Dual transaction with Accelerate Resources (19 May 2026 release) Kanowna East exploration integration FY2027 Acquired tenements to be incorporated into Maritana's FY2027 exploration strategy (19 May 2026 release) Black Swan FEED studies and Final Investment Decision (FID) Q2 CY2026 Per published Strategic Plan (April 2026 Investor Presentation) Black Swan plant construction Q3 CY2026 – Q2 CY2027 Per published Strategic Plan (April 2026 Investor Presentation) Black Swan first gold production Mid-2027 Per published Strategic Plan and Scoping Study (April 2026 Investor Presentation) Figure 6: Key Milestone and Timing for Black Swan Processing Hub (Source: Maritana Minerals Investor Presentation, April 2026) About Maritana Minerals Limited Maritana Minerals Limited (ASX: MRT) is a Western Australian gold developer positioning to become a standalone gold producer in the Kalgoorlie–Coolgardie region. The company holds a land package of approximately 1,167 square kilometres across the WA Goldfields and a Group Mineral Resource of 1.9 million ounces of gold (34.3Mt at 1.7g/t Au) across multiple deposits, including the cornerstone Burbanks and Boorara projects. The centrepiece of the company's strategy is the 2.2Mtpa Black Swan Processing Hub, located approximately 50 kilometres north-east of Kalgoorlie. The plant is currently in care and maintenance and is being refurbished and converted to gold processing, with the first gold targeted for mid-2027. Studies completed in February 2026 support the development of a gold operation producing more than 100,000 ounces per annum from the refurbished facility through a hub-and-spoke operating model. Maritana also holds the Nimbus Silver-Zinc deposit, located approximately 44 kilometres from Black Swan, which contains a Mineral Resource of 20.2 million ounces of silver, 104,000 tonnes of zinc and 77,000 ounces of gold and provides additional optionality across the portfolio. The company reported a market capitalisation of approximately A$332.8 million as at 14 April 2026, cash and cash equivalents of A$114.1 million as at 6 March 2026, and no debt. Maritana is fully funded for the refurbishment of Black Swan and the commencement of mining and gold production. Previous Samso News Coverage Samso has followed Maritana Minerals (formerly Horizon Minerals) across multiple ASX releases. The following represents our prior published coverage of the company: May 5, 2026 - October 8, 2025 - September 2, 2025 - August 21, 2025 - August 1, 2025 - July 27, 2025 - June 29, 2025 - Samso Concluding Comments The Maritana Minerals share price journey(Figure 7) is taking a decline over the last few months and this is not necessarily a reflection in the performance of the company but rather a reflection of the devaluing of the gold price (Figure 8), I think. If you look at the leaders in the sector such as Northern Star Resources Limited (ASX: NST), they are not fairing well too, coming off a high of almost AUD $32 to a recent price of AUD $18.83. Westgold Resources Limited (ASX: WGX) is coming of a high of just over AUD $8.00 to a current price of AUD $4.96. Figure 7: The Maritana Minerals Limited share price journey as of 22 May 2026. (Source: Commsec) There is definitely a sense of uneasiness in the sector and companies such as Maritana Minerals who are in the development phase are going to be vulnerable. Does this mean that there are going to be downgrades to come, I think the signs are definitely there but we all know that the market is always adjusting and the key for shareholders and potential investors is timing. Personally, with all the years of looking at the market in this sector, producers and aspiring producers are nearly always the first to reflect upturns and downturns. The gold price "run" is only really from September 2026 (Figure 8) and the monetary adjustments has only started. Are we going to see gold at thighs of over USD $5,000 again, I think this is a real possibility. When is that going to happen ? That is the million dollar question. We are starting to see the lithium price coming back. Lithium companies are restarting projects and the optimism in the air is getting moving. Nickel price which has been in the lows are starting to get headlines again. As we all know already, inflationary concerns are still the main topic of conversation and monetary policy controls are still very well alive. Figure 8: The gold price chart as of 25th May 2026. (Source: Trading Economics) In regards to companies like Maritana who have recently raised funds to proceed with development, this could well be a stocktake sale that serious investors should take seriously. Remember that investors tend to sell with a falling share price and buy on a rising tide. Maritana may not be a Northern Star at the moment but it definitely is a company that has gold to sell. Caution is the key word at the moment as Maritana is trading at a market capitalisation of just over AUD $260M as of 25th May 2026. I am sure that a business that has gold to sell when a resurgence of the gold price reappears, will be wroth more than its current level. The key is understanding of where that curve in Figure 8 is going to go in the next six months. Are we going to see more discounting and if so, is a series of accumulating the stock at a discounting price a good way or positioning yourself in a potential future gold producer. Cover photo: Gold panning on Jack Wade Creek (24785190007), BLM Alaska. Public domain, via Wikimedia Commons (https://commons.wikimedia.org/wiki/File:Gold_panning_on_Jack_Wade_Creek_(24785190007).jpg) The Samso Way – Seek the Research Here at Samso, we pride ourselves on delivering content for investors that is independent and informed by over three decades of experience in the industry. Our content is well-researched and is only created if I see merit in discussing the company's story. Our mission is simple: cut through the noise and spotlight what matters—genuine stories, grounded insights, and real opportunity. Our content is well-researched and is only created if the team sees merit in discussing the company or concept. Investors can explore our three core platforms: Coffee with Samso Samso Insights Samso News There may be numerous paths to success in investing, but the common thread among successful individuals is that they remain committed to making informed decisions. Equip yourself with the right knowledge and tools, and you will be well on your way to achieving your financial goals. Most importantly, investors need to be absolutely diligent in understanding their own risk-reward tolerance and capabilities. Never bite off more than you can chew. As they say, Rome wasn’t built in a day, and the Great Wall stood because it took centuries to complete. The Samso Philosophy: Stay curious. Stay sharp. And remember—digging deeper always uncovers the real value. In Life, there is no such thing as a Free Lunch. Never bite off more than you can chew is my parting comment. Happy Investing, and the only four-letter word you need to know is DYOR. To support our independent nature of our work, please head over to our Support Page and give us a helping hand in any of the ways listed. This is a new initiative for the Samso Platform, and it was always the concept of Samso when we started this journey in 2018. Disclaimer The information or opinions provided herein do not constitute investment advice, an offer, or solicitation to subscribe for, purchase, or sell the investment product(s) mentioned herein. It does not take into consideration, nor have any regard to your specific investment objectives, financial situation, risk profile, tax position and particular, or unique needs and constraints. Read full Disclaimer. Share to Grow: Your Bonus Samso has just released an eBook: How to Add Value to your Share Portfolio |A lesson on geological models sought by mining companies that gives insight and an understanding of which portfolios are better - and potentially more lucrative – investments. Click here to download this eBook.| Download eBook If you find this article informative and useful, please help me share the information. I try to write about topics that are interesting and have the potential to be of investment value. It is not easy to find stories that fit those parameters. If you or your organisation sees the benefit of what Samso is trying to achieve and has a need to share your journey, please contact me at noel.ong@samso.com.au. About Samso Samso is a trusted platform that equips dedicated investors with up-to-date industry knowledge and insigh0ts from top CEOs and thought leaders. By staying informed on business advancements and market trends, investors can enhance their financial decisions through a combination of expert guidance and their own research. Samso News | www.samso.com.au | An Investor Lens on ASX-Listed Companies

Samso Market Update | Monday, 19 October 2026 | Market close 4:30pm AEDT The S&P/ASX 200 closed up 39.0 points or 0.44% at 8,845.7 on the ASX today. Cleanaway Waste Management rose 20.2% and Challenger 19.3%, each on a takeover proposal announced before the open. The All Ordinaries, which tracks the 500 largest listed companies, rose 0.39% to 8,975.7. The S&P/ASX Small Ordinaries, an index of the companies ranked 101 to 300 by size, fell 0.35% to 3,177.9. This morning's pre-market edition is below, and Friday's editions are at ASX today, Friday 16 October 2026. ASX today: Key Highlights The ASX 200 rose 0.44% to 8,845.7. Energy was the best sector, up 2.21%, and information technology the worst, down 1.31%. Cleanaway (ASX: CWY) rose 20.2% to A$2.97 on the day it agreed to a A$3.095 a share cash takeover by EQT Infrastructure. Challenger (ASX: CGF) rose 19.3% to A$11.75 on the day it announced a non-binding A$11.80 a share proposal from Daiichi Life. Coal miners rose. Yancoal gained 5.7%, Whitehaven Coal 5.2% and New Hope 5.8%, which announced it will buy Malabar Resources. More shares fell than rose. Of the 1,500 companies that traded and have a market value on the ASX, 490 rose, 639 fell and 371 were unchanged. China's economy grew 4.5% in the September quarter from a year earlier, against a 4.4% forecast, the ABC reported. Takeovers, sectors and Bendigo Bank Five takeover and acquisition announcements were made before the open. A scheme of arrangement is a takeover that shareholders vote on and a court must approve. Cleanaway, FleetPartners and RAS Technology signed binding schemes. Challenger received a non-binding proposal, which is an offer that is not yet a firm deal. New Hope agreed to buy the 74% of unlisted Malabar Resources it does not own. At 12:28pm the ABC reported that energy, banks and takeover activity, naming Challenger and Cleanaway, were lifting the market. Energy rose 2.21%, the best of the 11 sectors (Yahoo Finance data). Market Index reported at 2:00pm that coal miners were "sharply higher". Separately, it reported at 1:48pm that China's coal output fell 6.8% from a year earlier in September. Information technology fell 1.31%. Market Index reported at 11:43am that the sector fell 2.5% following the Nasdaq's 2.1% fall on Friday. NextDC (ASX: NXT) entered a trading halt at 9:33am AEDT. It said it needs to give a material update on the impact of a delay to power supply at its M3 Melbourne data centre, including on FY27 guidance. Bendigo and Adelaide Bank (ASX: BEN) fell 4.0% to A$10.13 on the day of its first quarter trading update. The ABC quoted J.P. Morgan's Andrew Triggs, a bank analyst, who wrote that the result "looks a little soft on NIM/revenue". Blue Chips (above A$1 billion) Samso's four size bands use market value at Friday's close. They are not the S&P/ASX indices of the same names. Gainers. Cleanaway Waste Management (ASX: CWY) rose 20.24% to A$2.97 on turnover of A$625 million. It signed a scheme with Sedum Group, owned by funds managed or advised by EQT, at A$3.095 cash a share less any dividends paid. Cleanaway said that implies an enterprise value of about A$9.3 billion. The shares closed 12.5 cents below the A$3.095 offer price. Benz Mining (ASX: BNZ) rose 19.46% to A$4.85 on the day it reported more drill results from its Icon gold project, including 46 metres downhole at 6.5 grams of gold a tonne from 454 metres. Challenger (ASX: CGF) rose 19.29% to A$11.75. Daiichi Life, which owns 20.5% of Challenger, proposed to buy the rest at A$11.80 a share. The proposal is subject to due diligence and other conditions, and Challenger's board is considering it. FleetPartners Group (ASX: FPR) rose 7.13% to A$5.11 on the day it signed a A$5.30 cash a share scheme with SG Fleet. Ingenia Communities (ASX: INA) rose 7.07% to A$4.24, with no ASX announcement since Friday afternoon. New Hope Corporation (ASX: NHC) rose 5.82% to A$6.18. It will issue 0.5 of its shares for each Malabar share, under a scheme that Malabar shareholders vote on. Fallers. Mesoblast (ASX: MSB) fell 11.06% to A$1.77 on turnover of A$17.9 million, with no ASX announcement since Friday afternoon. Tasmea (ASX: TEA) fell 6.90% to A$9.31. Market Index reported at 1:01pm that engineering stocks were lower, naming Tasmea. Its annual report was released at 3:38pm. 4DMedical (ASX: 4DX) fell 5.38% to A$3.87 and Infratil (ASX: IFT) 5.28% to A$9.87. 4DMedical's only ASX announcement since Friday afternoon was a share quotation notice. Infratil had none. Stockland (ASX: SGP) fell 4.02% to A$4.06 on the day of its first quarter operational update and annual meeting. Mid Caps (A$500 million to A$1 billion) Gainers. Tuas (ASX: TUA) rose 5.18% to A$1.725, with no ASX announcement since Friday afternoon. Amplitude Energy (ASX: AEL) rose 3.88% to A$1.875 on the day it released its first quarter report. Nanosonics (ASX: NAN) rose 3.85% to A$2.97 and St Barbara (ASX: SBM) 3.77% to A$0.825. Neither had an ASX announcement since Friday afternoon other than a buy-back update from Nanosonics. Fallers. Weebit Nano (ASX: WBT) fell 7.64% to A$2.66 on the day it released an investor update. Market Index named it among the technology fallers. Deep Yellow (ASX: DYL) fell 6.10% to A$0.77, Alpha HPA (ASX: A4N) 5.56% to A$0.425 and Elevra Lithium (ASX: ELV) 5.49% to A$3.79. SKS Technologies Group (ASX: SKS) fell 5.07% to A$6.56. None of the four had a price-sensitive ASX announcement since Friday afternoon. Small Caps (A$150 million to A$500 million) Gainers. Wrkr (ASX: WRK) rose 12.5% to A$0.09 on turnover of A$146,000. Astron (ASX: ATR) rose 9.48% to A$0.635 on turnover of A$54,000, and Sky Metals (ASX: SKY) 9.38% to A$0.175. Dateline Resources (ASX: DTR) rose 9.21% to A$0.083 on turnover of A$1.0 million. None of the four had an ASX announcement since Friday afternoon. Fallers. Stakk (ASX: SKK) fell 20.83% to A$0.019, from A$0.024, on turnover of A$2.3 million. It announced that its agreement with Chime Financial, a material client, was extended to 31 December 2027 on unchanged terms. Tetratherix (ASX: TTX) fell 11.55% to A$5.36 and Nova Minerals (ASX: NVA) 11.24% to A$0.395. Cobre (ASX: CBE) fell 10.53% to A$0.255 on the day it named a new chair. Terra Metals (ASX: TM1) fell 10.0% to A$0.135, with no ASX announcement since Friday afternoon. Samso covered Terra Metals and its Southwest discovery in September. Micro Caps (below A$150 million) Gainers. RAS Technology Holdings (ASX: RTH) rose 85.19% to A$1.00. It signed a scheme with The Press Association, part of the UK's PA Media Group, at A$1.04 cash a share, a 93% premium to Friday's close of A$0.54. AXP Energy (ASX: AXP) rose 66.67% to A$0.005, from A$0.003, on turnover of A$149,000. In reply to an ASX price query, a letter asking a company to explain an unusual price move, AXP said it was not aware of any information that could explain the trading. Skylark Minerals (ASX: SKM) rose 19.51% to A$0.245. It reported that the Mineral Resource Estimate at its Zaranou gold project in Côte d'Ivoire rose 228% to 1.20 million ounces. A Mineral Resource Estimate is a company's measure of how much metal it believes is in the ground. Eminence Minerals (ASX: EMA) rose 15.56% to A$0.052 on the day it secured exclusivity over a Côte d'Ivoire gold portfolio. Alvo Minerals (ASX: ALV) rose 12.79% to A$0.097 on turnover of A$4.9 million, with no ASX announcement since Friday afternoon. Fallers. Infinity Mining (ASX: IMI) fell 28.57% to A$0.005 on turnover of A$70,000, with no ASX announcement since Friday afternoon. At prices this low, a move of a fraction of a cent is a large percentage. Albion Resources (ASX: ALB) fell 14.52% to A$0.053 on the day it announced a A$1.8 million placement to fund gold exploration at Gidgee. Vertex Minerals (ASX: VTX) fell 11.82% to A$0.097 on the day of an operational update on its Reward gold mine. Black Bear Minerals (ASX: BKB) fell 11.24% to A$0.395 and Australian Vanadium (ASX: AVL) 10.53% to A$0.17, with no price-sensitive ASX announcement since Friday afternoon. Commodities and Currency Prices between 4:21pm and 4:33pm AEDT (Yahoo Finance data), compared with the figures in this morning's edition. Front-month gold futures were at US$4,196.40 an ounce, up from US$4,181.10. Front-month copper futures were at US$6.61 a pound, up from US$6.57. Brent crude futures were at US$108.00 a barrel, down from US$109.86. Iron ore was last reported at US$91.20 a tonne on Friday, up 1.5% (ABC News). Samso found no Monday price by the close. The Australian dollar was buying 69.76 US cents at 4:32pm AEDT, up from 69.63 US cents at 8:44am AEDT. Samso Concluding Comments The ASX 200 rose 0.44% to 8,845.7. Five takeover and acquisition announcements were made before the open, and Cleanaway and Challenger each rose about 20%. Energy led the sectors, technology fell, and 639 companies fell against 490 that rose. Cleanaway, Challenger and energy shares rose while most shares and the Small Ordinaries fell. NextDC remains in a trading halt until its update on the M3 power delay, due by Wednesday's open at the latest. General information only. This is not financial advice. Samso Market Update | Monday, 19 October 2026 | Pre-market 9:45am AEDT Futures point to a slightly lower open on the ASX today. ASX 200 futures were down 0.1% at 8,826 points at about 7:00am AEDT, according to the ABC News market snapshot. A futures contract is an agreement to buy or sell at a set price on a later date, and its price is used as a guide to where the index will open. The 0.1% fall is measured on the futures contract, not against Friday's index close. The S&P/ASX 200 closed on Friday at 8,806.7, up 87.0 points or 1.00%. The All Ordinaries, which tracks the 500 largest listed companies, rose 1.13% to 8,940.4. The S&P/ASX Small Ordinaries, an index of the companies ranked 101 to 300 by size, rose 2.20% to 3,189.2. Friday's editions are at ASX today, Friday 16 October 2026, and Thursday's are at ASX today, Thursday 15 October 2026. ASX today: Key Highlights ASX 200 futures were down 0.1% at 8,826 at about 7:00am AEDT (ABC News). The Nasdaq Composite fell 2.15% and the S&P 500 fell 1.09% on Friday. The Dow Jones Industrial Average fell 0.34%. Brent crude futures settled at US$109.86 a barrel, up 3.7%, the highest settlement since May according to FinanceFeeds. The Nikkei 225 rose 1.39% on Friday but fell 5.37% over the week, according to Trading Economics. China releases its third-quarter growth figures today, according to the ABC. The market forecast is 4.4% (Seoul Economic Daily). Wall Street overnight Wall Street finished lower on Friday. The Dow Jones Industrial Average fell 175.28 points or 0.34% to 52,002.41. The S&P 500 fell 84.62 points or 1.09% to 7,659.44. The Nasdaq Composite fell 573.00 points or 2.15% to 26,140.03. The Russell 2000, an index of smaller US companies, fell 1.66% to 2,721.37. The ABC reported "growing anxiety of the debt financing of the AI boom and whether the massive investment will deliver decent returns." NAB senior FX strategist Rodrigo Catril, quoted by the ABC, called the fall "a reassessment of the economics underpinning the AI investment boom." The ABC reported that CoreWeave, which runs data centres, fell 12% on Friday and about 25% over the week. TheStreet reported that energy shares rose 2.52% with the rise in oil prices. The US 10-year Treasury yield was at approximately 5.10% late in the New York session, up about 1 basis point from Thursday (Yahoo Finance data). A Treasury is a US government bond. A bond yield is the yearly return from buying a government bond at today's price and holding it until it is repaid, and when yields rise, bond prices fall. A basis point is one hundredth of a percentage point. Japan The Nikkei 225 closed on Friday at 65,324.29, up 898.34 points or 1.39%. Trading Economics reported that selling of technology shares eased, after TSMC reported a record third-quarter profit and Nvidia chief executive Jensen Huang defended CoreWeave's financial health. It named Lasertec (up 5.1%) and Kioxia (up 3.8%) among the gainers. The index fell 5.37% over the week, its steepest weekly fall since July, according to Trading Economics. The US dollar was buying approximately 157.97 yen at 8:46am AEDT. Hong Kong's Hang Seng rose 0.52% and the Shanghai Composite rose 0.24% on Friday. London and Europe London's FTSE 100 closed on Friday at 10,881.90, up 41.50 points or 0.38% (Yahoo Finance data). Germany's DAX rose 0.95% to 25,622.61. France's CAC 40 rose 0.28% to 7,956.70. Currencies The Australian dollar was buying approximately 69.6 US cents at 8:44am AEDT, up 0.64% from Friday's session. The ABC quoted Barclays saying it was bullish on the Australian dollar and that "Australia remains a key beneficiary of the AI commodity supercycle." The US dollar index, which measures the US dollar against six major currencies, was at approximately 101.78 late in the New York session on Friday, down 0.22%. The euro was at approximately US$1.119. Commodities Brent crude futures for the front month, the contract that falls due soonest, settled on Friday at US$109.86 a barrel, up US$3.94 or 3.7%. Brent briefly touched US$110, its highest since May, according to the ABC. West Texas Intermediate, the US oil benchmark, settled at US$101.41 a barrel for November delivery, up 3.5% (ABC News, FinanceFeeds). FinanceFeeds reported that a Saudi-led coalition said it intercepted a ballistic missile aimed at the Red Sea export port of Yanbu, and that a Ukrainian drone strike shut a main unit at a large Russian refinery. Gold futures for December settled at US$4,181.10 an ounce, up 0.26% (Yahoo Finance data). Silver futures for December rose 0.82% to US$57.31 an ounce. Copper futures for December rose 0.95% to US$6.5735 a pound. The ABC reported that copper fell over the week as the end of a strike at a large Chilean mine eased supply concerns. Iron ore rose 1.5% to US$91.20 a tonne on Friday, according to the ABC News market snapshot (prices at about 7:00am AEDT). Today's diary All times AEDT. No Australian Bureau of Statistics releases are scheduled today (ABS release calendar). Reserve Bank Assistant Governor (Financial Markets) Christopher Kent is due to speak, according to the ABC. No time was given. 1:00pm. China's third-quarter gross domestic product (GDP), the total value of goods and services an economy produces in a period, with September retail sales, industrial production and fixed asset investment (Trading Economics calendar, 02:00 UTC). The market forecast for third-quarter growth is 4.4% (Seoul Economic Daily). Samso Concluding Comments ASX 200 futures were down 0.1% at about 7:00am AEDT. On Friday the Nasdaq fell 2.15%, with the ABC and NAB linking the fall to concern over the cost and returns of AI investment, and Brent crude rose 3.7% to US$109.86 a barrel. China's third-quarter growth figures at 1:00pm AEDT are the main scheduled item today. General information only. This is not financial advice.

Rooster Talk Episode 61 is with Matthew Boyes, Managing Director of Red Dirt Metals Limited (ASX: RDT). The Mt Ida project is fast becoming a double bonus for Red Dirt Metals as exploration is unveiling the gold credits. When we were first introduced to the Mt. Ida project (Red Dirt Metals Limited (ASX:RDT) - A Lithium Gem within a Gold Mine.) in April 2022, we were just trying to understand the lithium potential. As the market was having a lithium boom, it would be foolish not to discuss what the flavour of the last two years has been. As we were talking through that Coffee with Samso with Matt Boyes, I was really interested in the gold potential. The Mt Ida story is very similar to what occurred at Mt. Holland a few years ago. For those that have not heard the story, Mt Holland was all about the discovery of a lithium bonanza because of a phone call. Mt. Holland was a legitimate gold mining project but was later discovered to be an even greater lithium project. When you hear Matt talk about the lithium prospectivity in Mt. Ida, you can hear the frustration of not being able to rush into the gold project. The potential is yet to be finalised. In saying that, you can also feel the frustration of not being able to unwrap the potential of the lithium prospectivity. The fact that there is still a large package of landholding that the company has not explored is definitely high on the list of things to do. Matt is certain that there will be a resource coming out for the lithium and in time, there will be a gold resource to be released as well. When that will be is the million dollar question. Check out this video - this is definitely a good story to follow. Chapters: 00:00 Start. 00:20 Introduction. 01:07 Matthew updates. 02:38 Relationship between Gold and Lithium. 03:56 Mineralisation of Lithium and Gold. 06:41 Announcement key points for investors. 08:18 Understanding the geology from the announcement. 10:35 Future plans for the project. 11:59 Lithium market discussion. 15:24 Updates on covering the entire tenement package. 16:57 Potential focus on Gold. 19:49 Exploration potential. 21:32 Matthew summarises the news flow. 22:54 Conclusions. PODCAST This is a good time to download the first Ebook (FREE) from Samso as it is all about VMS (Volcanogenic Massive Sulfides). About Matthew Boyes Managing Director Mr Boyes is a qualified Geologist and Fellow of the Australasian Institute of Mining and Metallurgy with 25 years industry experience and a strong background in minerals resources, project development, construction and operations management. Mr Boyes has recently served as COO for AIM listed Patagonia Gold successfully overseeing the development and construction of two standalone Gold-Silver heap leach projects in southern Argentina. Mr Boyes has significant experience in product evaluations and financing with a focus on the junior market sector. About Red Dirt Metals Limited (ASX: RDT) Red Dirt is focused on the development and evaluation of underexplored brownfields exploration assets in Western Australia. Mt Ida Gold Pty Ltd is the holding company of the Mt Ida Gold Project, comprising 19 tenements in the historical Mt Ida gold mining district, located approximately 100km north-west of Menzies in the WA goldfields. The project tenements cover an area of 155km2 and include the historic Timoni Gold Mine and others, which have produced greater than 300,000 oz gold at 17.2 g/t head grade. Highlights: Acquired from Ora Banda Mining in September 2021 for $11M cash A current gold resource of 318,000 t @13.8g/t gold for 141koz calculated under the JORC 2004 guidelines Evidence of occurrences of Lithium-bearing pegmatites with Tantalum Exploration activities to commence immediately October 2021 The Mt Ida project has total resources of 318,000 t @13.8g/t of gold for 141koz, Indicated resources total 136,000t @ 18.6 g/t gold for 81koz with 182,000t @ 10.3 g/t gold for 60koz in Inferred category. All resource currently stated at Mt Ida are calculated under the JORC 2004 guidelines with no material change to the existing resources since estimation. Lithium Review: Whilst undertaking the technical due diligence of the Mt Ida Project, it was noted that multiple high-grade lithium intervals had been intersected within pegmatite bodies on the western contact of the main Mt Ida granite intrusive, known as the Copperfield Granite. Initial investigations from available data shows pegmatite outcropping in 5 separate locations proximal to the main granitic intrusive, and along the contact with the western bounding mafic amphibolite units which host the high-grade Mt Ida gold copper lodes. A preliminary target corridor has been interpreted and extends for up to 5 km in strike from the most southern-identified outcrop parallel to the granite and north into the Company’s adjacent exploration tenement, with potential to extend further in both directions. Of significance is that no historic exploration has been carried out with the intention of targeting these Lithium bearing pegmatite intrusives. This opens up large areas on both the eastern and western limbs of the granitic intrusive contacts for future exploration and will complement the Company’s aggressive gold/copper exploration program at the Mt Ida Project. Disclaimer The information or opinions provided herein do not constitute investment advice, an offer or solicitation to subscribe for, purchase or sell the investment product(s) mentioned herein. It does not take into consideration, nor have any regard to your specific investment objectives, financial situation, risk profile, tax position and particular, or unique needs and constraints. Read full Disclaimer. About Samso Keep us informed too! Please let us know your thoughts and send us any comments to info@samso.com.au. Remember to Subscribe to our YouTube Channel, Samso Media and our mail list to stay informed and make comments where appropriate. Other than that, you can also give us a Review on Google.

Coffee with Samso Episode 135 is with Matthew Boyes, Managing Director of Red Dirt Metals Limited (ASX: RDT). I first spoke to Matthew Boyes, Managing Director of Red Dirt Metals Limited (ASX: RDT) in May 2021 when the company was known then as TNT Mines Limited. That was the first Coffee with Samso for Matthew and we spoke about the direction of the company then. Check out the Coffee with Samso: The Remaking of the Eureka Gold Mine - TNT Mines Limited (ASX: TIN) - Episode 86 Since the acquisition of the Mt Ida project in September 2021 (Mt Ida Acquisition, Capital Raise and Presentation), the company has rebranded themselves and is now on the greater path of developing a Lithium project. The company completed the transaction on 24th September 2021 (Completion of Mt Ida Acquisition). In this episode of Coffee with Samso, Matthew gives us an insight into the story of Mt. Ida. The movement in share price and hence, market capitalisation has been remarkable. However, at the moment of writing this, the value of AUD212M according to Matthew, seems to be undervalued. The project brings back recollection of the Mt Holland story, but time will tell if the Mt. Ida lithium story will bring similar fortunes for shareholders of Red Dirt Metal. Chapters: 00:00 Start 00:20 Introduction 01:07 The Mt Ida Story 05:28 What kind of grade does hard rock lithium project need in relation to Red Dirt? 08:36 Advantages of drilling for pegmatites. 10:19 How much of the project has been drilled? 12:22 Can you see consistencies in the geology of the pegmatites? 14:01 What should people look for to understand the Red Dirt Story? 15:32 Explaining the significance of LCT pegmatites. 17:22 Are the proportions of LCT positive or negative? 19:02 What could throw a spanner in the works? 20:21 Is there a type deposit happening in the Lithium industry? 22:32 How does Matthew take this Lithium market? 24:06 Is there such a term as a good or bad lithium project? 25:51 How should investors temper their expectations on the Red Dirt Story? 27:21 The lack of information is mutual. 28:21 Mt Ida is a Gold project too. 31:19 How does the gold sit spatially to the lithium? 32:56 Where does the Gold project sit on the list of things to do? 33:43 News flow. 35:16 What are the brokers saying about RDT? 36:04 How big can Mt Ida be? What is your pitch to Investors out there? 37:36 Samso's thoughts. 38:10 Conclusions PODCAST This is a good time to download the first Ebook (FREE) from Samso as it is all about VMS (Volcanogenic Massive Sulfides). About Matthew Boyes - Managing Director Mr Boyes is a qualified Geologist and Fellow of the Australasian Institute of Mining and Metallurgy with 25 years industry experience with a strong background in minerals resources, project development, construction and operations management. Mr Boyes has recently served as COO for AIM listed Patagonia Gold successfully overseeing the development and construction of two standalone Gold-Silver heap leach projects in southern Argentina. Mr Boyes has significant experience in product evaluations and financing with a focus on the junior market sector. About Red Dirt Metals Limited (ASX: RDT) Red Dirt is focused on the development and evaluation of underexplored brownfields exploration assets in Western Australia. Mt Ida Gold Pty Ltd is the holding company of the Mt Ida Gold Project, comprising 19 tenements in the historical Mt Ida gold mining district, located approximately 100km north-west of Menzies in the WA goldfields. The project tenements cover an area of 155km2 and include the historic Timoni Gold Mine and others, which have produced greater than 300,000 oz gold at 17.2 g/t head grade. Highlights: Acquired from Ora Banda Mining in September 2021 for $11M cash A current gold resource of 318,000 t @13.8g/t gold for 141koz calculated under the JORC 2004 guidelines Evidence of occurrences of Lithium-bearing pegmatites with Tantalum Exploration activities to commence immediately October 2021 The Mt Ida project has total resources of 318,000 t @13.8g/t of gold for 141koz, Indicated resources total 136,000t @ 18.6 g/t gold for 81koz with 182,000t @ 10.3 g/t gold for 60koz in Inferred category. All resource currently stated at Mt Ida are calculated under the JORC 2004 guidelines with no material change to the existing resources since estimation. Lithium Review: Whilst undertaking the technical due diligence of the Mt Ida Project, it was noted that multiple high-grade lithium intervals had been intersected within pegmatite bodies on the western contact of the main Mt Ida granite intrusive, known as the Copperfield Granite. Initial investigations from available data shows pegmatite outcropping in 5 separate locations proximal to the main granitic intrusive, and along the contact with the western bounding mafic amphibolite units which host the high-grade Mt Ida gold copper lodes. A preliminary target corridor has been interpreted and extends for up to 5 km in strike from the most southern-identified outcrop parallel to the granite and north into the Company’s adjacent exploration tenement, with potential to extend further in both directions. Of significance is that no historic exploration has been carried out with the intention of targeting these Lithium bearing pegmatite intrusives. This opens up large areas on both the eastern and western limbs of the granitic intrusive contacts for future exploration and will complement the Company’s aggressive gold/copper exploration program at the Mt Ida Project. Disclaimer The information or opinions provided herein do not constitute investment advice, an offer or solicitation to subscribe for, purchase or sell the investment product(s) mentioned herein. It does not take into consideration, nor have any regard to your specific investment objectives, financial situation, risk profile, tax position and particular, or unique needs and constraints. Read full Disclaimer. About Samso Keep us informed too! Please let us know your thoughts and send us any comments to info@samso.com.au. Remember to Subscribe to our YouTube Channel, Samso Media and our mail list to stay informed and make comments where appropriate. Other than that, you can also give us a Review on Google.

Japan spent three decades in near-zero growth after its property bubble burst in 1991, and only in the last two years has a genuine recovery finally taken hold. China's property bubble has been bursting since 2021, its debt is harder to see clearly than Japan's ever was, and Beijing has already legislated against a threat, AI-driven job losses, that Washington is still only debating. What the gap between the two economies means for where the next decade of ASX-relevant growth is likely to land is the question this piece sets out to answer. Samso Insights Macro & Geopolitical China AI & Capital READING THE SERIES This piece runs in two parts. Part 1, this piece. Sets out the macro case: Japan's precedent, China's property crisis, its debt structure, the population decline and the AI-jobs tension arriving at the same time. Part 2. Looks at where the capital is actually flowing, and the ASX and TSX-listed tungsten, tin, copper and iron ore producers on the other side of that trade. WHAT THIS PIECE COVERS 1.00 THE ONLY PRECEDENT ON RECORD What happened to Japan after 1991, and why it is the only comparison available. 2.00 THE MEGACITY PLAN The 2013 plan to move 250 million people into cities, and what has replaced it. 3.00 THE PROPERTY CRISIS Where the developer collapse now stands, legally and economically. 4.00 THE UNFINISHED CITIES How much was overbuilt, and why nobody agrees on the number. 5.00 TWO WAYS TO OWE MONEY Japan's debt is measured. China's is a range, and the range is the problem. 6.00 THE POPULATION THAT ISN'T COMING The cities are growing while the country is shrinking. 7.00 AI ARRIVES AT THE SAME MOMENT What Beijing is legislating against, and the jobs question underneath it. 8.00 WHAT THE FORECASTS ACTUALLY SAY A managed slowdown, with the weak point named by the forecasters themselves. 9.00 WHERE THIS LEAVES THE READER The slower path, the harder path, and what decides which one China gets. A note on the vocabulary and the sources. Technical and mining terms are explained in plain English where they first appear, and all of them are collected in one box at the back. Every source carries an identifier, [R1] to [R50], and that identifier appears in the body text at the point the source is used, so any figure can be traced to where it came from. 1.00 — THE ONLY PRECEDENT ON RECORD Japan already ran this experiment once In 1991, Japan's property bubble burst, and the country spent the next three decades in near-zero growth, near-zero inflation, and a debt load that climbed past 200% of GDP the entire time. It survived intact, no default, no collapse, and only in the last two years has a genuine recovery finally taken hold, with sustained inflation, real wage growth, and interest rates rising for the first time since the 1990s. Japan is the closest thing on record to a completed experiment in how a debt-fuelled property bubble plays out over thirty years (see FIG. 01). FIG. 01 — Japan's urban land price index, six largest cities, residential, indexed to the 1991 peak. Source: Japan Real Estate Institute data as reported by CEIC, on the index level through 2017 (the most recent year with a directly comparable published figure on this index series). *The 2026 point is Samso's own estimate, not a directly published index value on this series, based on Japan's Ministry of Land, Infrastructure, Transport and Tourism reporting nationwide land prices rising for a fifth consecutive year to March 2026, the strongest annual increase since 1992.[R37] China's property bubble started bursting in 2021, and by most of the measures that matter, investment, construction, sales, it is still falling. That alone would make the Japan comparison worth drawing. What sharpens it is that China is also carrying two pressures Japan never had to face at the same time, a population that is already shrinking rather than merely ageing, and an AI-driven threat to jobs that barely existed in any serious form during Japan's own lost decades. Beijing has already legislated against that threat, banning AI-driven dismissals and proposing automation caps, while Washington is still only debating whether to. This connects beyond China itself. China is now running the live version of the experiment Japan already completed, with fewer of Japan's advantages and at least two extra variables layered on top. Where that experiment lands, and how long it takes, matters directly to anyone trying to work out where the next decade of ASX-relevant growth is likely to come from. That's the question this piece, and Part 2 after it, is built to answer. 2.00 — THE MEGACITY PLAN Renewal has replaced expansion, but the assumptions haven't changed The idea that China would move 250 million rural residents into cities dates back to a 2013 State Council plan.[R1] The goal then was to get 70% of the population, roughly 900 million people, living in cities by 2025, up from about half at the time. The logic rested on one idea. Rural households in China are largely self-sufficient and consume little, so moving them into cities, and giving them city incomes and city habits, would create a domestic consumer base that doesn't depend on exports. That plan is now over a decade old, and China has moved on to a different phase. The 15th Five-Year Plan, covering 2026 to 2030, has shifted the language from expansion to renewal.[R2] The urban renewal campaign under that plan is expected to require at least 15 trillion yuan, around US$2.2 trillion, over five years, covering everything from pipelines and drainage to dilapidated housing and elderly-care facilities.[R3] Separately, Beijing has outlined a plan to fuse major cities into 19 megaregions, of which Hong Kong's Northern Metropolis project, linking into the wider Greater Bay Area, is one example already under construction (see FIG. 02).[R4] FIG. 02 — Shenzhen, seen from Hong Kong across the border, two of the cities being fused into the Greater Bay Area megaregion. Photo: Sparktour, Wikimedia Commons, 29 August 2019, CC BY-SA 4.0.[R28] The mechanism meant to make this work is reform of the hukou, China's household registration system.[R5] Historically, your hukou tied your access to schooling, healthcare and welfare to wherever you were officially registered, not wherever you actually lived, which is why migrant workers who moved to cities for decades often couldn't access the services those cities provided. Under the current urbanisation plan, rural residents face almost no restrictions moving to any city under 5 million people, and are promised the same entitlements as existing residents. That answers the access question. It doesn't fully answer the income question. Having the legal right to use a city service and having the income to make use of it are different things, and the plan is considerably stronger on the first than the second. There's also a quieter admission buried in the policy detail. Land supply is now meant to be linked to actual population flows. Cities gaining population get more land allocation, and cities losing population get less, or none. The Peterson Institute for International Economics, a credible source on this, notes that the government has effectively acknowledged there are two separate property markets in China.[R6] One is undersupplied, in top-tier cities with genuine housing affordability problems. The other is oversupplied, in lower-tier cities, some of which have become what's commonly called "ghost towns," built but never filled. That cuts against a simple story of Beijing building megacities and moving low-income households into them. What's actually happening is closer to Beijing trying to stop building in the places nobody wants to go, while concentrating what's left into a smaller number of winners. This mechanism has a name outside China too, and a much longer history than the current policy suggests. Western planners call it "rightsizing" or "shrinking cities" planning, and it has been standard practice in the United States and Germany for over twenty years, adopted for the identical fiscal reason.[R32] Youngstown, Ohio, drafted its "Youngstown 2010" plan in 2002, explicitly to stop building and maintaining infrastructure for a population that had fallen from 170,000 in 1930 to a fraction of that (see FIG. 03). Detroit, Flint, Buffalo and Cleveland followed with similar plans. Germany ran a national program called Schrumpfende Städte, shrinking cities, doing the same thing across the former East Germany after reunification, when large numbers of people left for the west. FIG. 03 — Youngstown, Ohio, whose population fell from around 170,000 in 1930 to a fraction of that, prompting the city's "Youngstown 2010" rightsizing plan in 2002, twenty years before China's own version of the same idea. Photo: Dblcut3, Wikimedia Commons, 6 August 2020, CC BY-SA 4.0.[R40] The core idea, concentrating services and infrastructure spending where people actually live rather than where they used to live, is identical in mechanism to what China's land-supply-follows-population-flow policy is doing. Beijing is applying a fiscal-discipline idea Western cities adopted decades ago, under very different political systems, for the same underlying reason. No government can afford to maintain infrastructure indefinitely for people who have already left. 3.00 — THE PROPERTY CRISIS A legal chapter closed in August. The economic one hasn't The urban renewal spending comes on top of a property sector that is still working through the largest developer crisis in China's modern history. Real estate's share of China's GDP fell to about 19% in 2024, down from a peak of 24% in 2018, according to Bloomberg Economics estimates.[R7] This is China's version of what happened to Japan in 1991, a debt-fuelled property boom running into its own weight, though it arrived three decades later and with a much larger economy underneath it. The headline event closed out this year. On 20 August 2026, a Shenzhen court sentenced Evergrande founder Hui Ka Yan to life imprisonment for financial crimes, and ordered the confiscation of his personal assets, following a guilty plea earlier in the year covering illegal fundraising, financial fraud, misuse of funds and bribery.[R8] A Guangzhou court has since accepted a bankruptcy liquidation case against Evergrande's mainland property development unit, which carried much of the group's roughly US$300 billion in total liabilities. Evergrande's own default in 2021 is the event that started the wider crisis, and its stock plunge dragged down other major developers including Country Garden and Kaisa (see FIG. 04). FIG. 04 — Evergrande-developed residential towers, still branded with the company's name, of the kind built across dozens of Chinese cities during the boom years. Photo: Windmemories, Wikimedia Commons, 5 March 2021, CC BY-SA 4.0.[R41] But the sentence closes a legal chapter, not an economic one. Property investment fell 19.2% in the first seven months of 2026, new construction starts dropped 24%, and sales of newly built commercial property fell 11.8% by floor area, according to Bloomberg reporting.[R9] Beijing has been easing its way out of the crisis rather than letting the market clear it on its own terms. In January 2026 it dropped the "three red lines" policy, the balance-sheet rules on developer borrowing that were originally what triggered Evergrande's collapse when they were introduced in 2020, and it's now leaning the other way, extending credit to help developers finish stalled projects.[R10] The state is trying to do two things that pull in different directions at once. It's winding down oversupply built during the boom years, while building new capacity in the cities it wants to grow. Whether the same low-income households the urbanisation plan is meant to serve can actually move between those two categories of city, from the ones losing population and land allocation to the ones gaining both, is the practical question the policy documents don't fully resolve. Samso Access Some of the work does not appear here Samso holds back a set of Insights for members. Same method as this piece, the same sourcing and the same counterweights, on companies and sectors that do not make the public feed. Every member-only Samso Insight unlocked Full access to the gated Insights archive New member-only pieces included as they publish Samso News, Coffee with Samso and Research Workups will join the gated set later. A$20 a month · cancel any time See what is inside → 4.00 — THE UNFINISHED CITIES Nobody, including Beijing, agrees on how much was overbuilt The oversupply mentioned above is physical. Rows of finished but empty apartment towers stand in one category, and apartments sold years ago and never finished at all stand in another. Nobody, including China's own statistics apparatus, has an agreed number for how much of this exists. He Keng, the former deputy head of China's National Bureau of Statistics, said at a public event reported by Reuters that estimates of the country's vacant housing stock vary wildly between experts, with some of the more extreme estimates suggesting there's enough vacant housing for 3 billion people.[R11] His own, more conservative assessment was that "1.4 billion people probably can't fill them," meaning the entire population of China, if every person needed a separate home, still wouldn't use up what's already built. The fact that a former senior statistics official is comfortable saying the real number is unknown tells you something about how large and how uncertain this problem is. One figure traces to a single, well-documented case. The Wall Street Journal has reported that of the roughly 1.2 million residential units Evergrande presold to buyers before construction was finished, about 800,000 remain unfinished.[R12] Those are homes people paid for, in some cases years ago, that still don't exist as habitable buildings (see FIG. 05). FIG. 05 — Unfinished apartment buildings in Maoming, Guangdong, representative of the stalled construction found across China's oversupplied lower-tier cities. Not a specific Evergrande site. Photo: 中少, Wikimedia Commons, 27 July 2021, CC BY-SA 4.0.[R29] Kenneth Rogoff, the Harvard economist, summed up the shape of it plainly on the Brookings Podcast in April 2026. China has "reached a classic overbuilding boom-bust," he said, and "it doesn't seem anywhere near the end."[R13] The single most widely reported case is Ordos, in Inner Mongolia, where the Kangbashi New Area was built with capacity for around 300,000 people and, years after completion, still housed only a fraction of that (see FIG. 06). FIG. 06 — The Ordos Museum in Kangbashi, Inner Mongolia, the single most widely reported example of a Chinese "ghost city," built for around 300,000 people with only a fraction of that number ever moving in. Photo: Popolon, Wikimedia Commons, 5 May 2015, CC BY-SA 4.0.[R43] This connects directly back to the urbanisation and renewal story above. Every yuan sunk into an apartment tower that never sold, or a development that was never finished, is capital that can't be recovered and can't be redirected into the megaregions Beijing now wants to build instead. It also helps explain why developers are in the state they're in, and why the state has had to step in with credit support rather than let the market clear the way it normally would. A construction boom that overbuilt this badly doesn't just waste money, it leaves the companies that did the building under too much financial strain to be reliable partners in the next phase of the plan. 5.00 — TWO WAYS TO OWE MONEY Japan's debt is enormous but known. China's is a range China's debt is often compared to Japan's, and the comparison is worth making properly, because the two situations look similar on the surface and are actually different underneath. This is also where the precedent from section 1.00 earns its keep. Japan's debt survived three decades of near-zero growth without a crisis for a specific, checkable reason, not luck, and whether China has the same protection is the real question behind this comparison. Japan carries the highest government debt load of any major economy, at roughly 204% of GDP by the IMF's 2026 estimate.[R14] That figure alone sounds alarming. What keeps it from being one is who holds it. Around 88% of Japan's government debt is held domestically, in yen, mostly by the Bank of Japan itself, which owns close to half of it outright, with most of the remainder held by Japanese banks and insurance companies. Foreign investors hold less than 10%. Japan's debt is recorded on a single, visible balance sheet, and the country that owes it and the institutions that hold it are, for the most part, the same country. That's the mechanism that let Japan wait out three decades of stagnation rather than default (see FIG. 07). FIG. 07 — The Bank of Japan's head office in Tokyo, the institution that holds close to half of Japan's government debt outright. Photo: yt_siden, Wikimedia Commons, 29 August 2011, CC BY-SA 2.0.[R38] China's debt is also mostly domestic and mostly denominated in its own currency, and that's significant, because it means China isn't exposed to the kind of crisis that hits countries which borrow heavily in foreign currency from foreign lenders, the pattern behind most emerging-market debt crises of the past forty years. On that specific point, China and Japan are more alike than different. Where they diverge is transparency. Japan's number is one number, and it's public. China's isn't, and nobody outside Beijing knows the real figure with confidence. The IMF estimated in 2025 that China's explicit local government debt reached 39% of GDP, around US$8 trillion, with a further 51% of GDP, about US$10.4 trillion, held in local government financing vehicles, known as LGFVs, entities that local governments set up specifically to borrow money outside the limits Beijing officially places on them.[R15] Add those two figures together and local government-linked debt alone approaches 90% of China's GDP, before the central government's own borrowing is counted. Other analysts, folding in a wider set of off-balance-sheet liabilities, put China's total public sector debt burden at 130% to 150% of GDP.[R16] S&P Global, assessing an earlier phase of the same problem, called China's hidden local debt "a debt iceberg with titanic credit risks."[R17] That range, a Fund estimate near 90% for local debt alone against outside analysts putting total public debt half again as high, is itself the point. Japan's debt is a single number. China's is a range. The risk also isn't spread evenly the way Japan's is. Guizhou, one of China's poorer provinces, carried LGFV debt close to 75% of its own provincial GDP, well past the 60% threshold Beijing has set as a warning line.[R18] Inner Mongolia, Ningxia, Liaoning and Qinghai are in similar territory. Wealthier coastal provinces are in a completely different position. China doesn't have one debt problem, it has dozens of provincial-level debt problems of very different severity, stitched together under one national figure that obscures how serious the worst of them actually are (see FIG. 08). FIG. 08 — A highway bridge under construction in Guizhou, the kind of large-scale infrastructure project financed through local government debt in one of China's most indebted provinces relative to its own GDP. Photo: Glabb, Wikimedia Commons, 21 August 2011, CC BY-SA 3.0.[R39] Beijing knows this and is working on it. A 12 trillion yuan, around US$1.7 trillion, package rolled out in late 2024 is aimed at resolving officially recognised hidden local government debt by mid-2027, and it has already brought that specific category down.[R19] The harder problem is the roughly 14.8 trillion yuan of LGFV operational debt that remains, much of it in high-interest, non-standard loans that are proving harder to unwind than the debt Beijing has already dealt with. Japan and China both largely owe themselves rather than foreign creditors. The real difference is that Japan's debt is enormous but known, recorded on one balance sheet everyone can see, while China's is opaque, fragmented across thousands of local vehicles, and concentrated unevenly in provinces already under the most strain from the property downturn and the slower urbanisation this piece has been describing (see TABLE 01). A debt problem you can see and price is a very different kind of risk to one where even the people managing it aren't fully sure of its size. TABLE 01 — Japan and China's government debt compared Japan China Headline government debt ~204% of GDP (IMF, 2026 est.) ~68% of GDP (2026 est.) to ~77% of GDP (IMF, 2022); two different sources at two different points in time, not one consistent measure Once local vehicles are included Not applicable, one consolidated figure Local debt plus LGFVs sum to approximately 90% of GDP, Samso's own addition of two separate IMF component figures (39% plus 51%, 2025); broader outside estimates run to 130 to 150% of GDP Domestic ownership ~88% domestically held; Bank of Japan alone holds close to half Mostly domestic and yuan-denominated, but the true scale is disputed Foreign ownership Under 10% Comparatively low, though harder to verify given data opacity Distribution of risk Uniform, one national figure Highly uneven; several provinces (Guizhou, Inner Mongolia, Ningxia, Liaoning, Qinghai) exceed Beijing's own 60% provincial risk threshold Coffee with Samso The same companies, on the record Coffee with Samso is the interview series. Management answering questions in long form, so the people behind a project can be judged alongside the numbers in a piece like this one. Free to watch. Companies commission an interview, and it stays free to everyone, permanently. Watch the interviews → 6.00 — THE POPULATION THAT ISN'T COMING The cities are growing. The country isn't This is the part of the story that turns three separate headlines into one argument. China's population peaked in 2022 and has fallen every year since.[R20] The country is projected to lose around 3.2 million people in 2026 alone, more than the entire population of the US state of Arkansas, with annual losses accelerating through the 2030s and 2040s under the United Nations' median projection. China's working-age population, defined as ages 16 to 59, fell from 61.3% of the total population in 2023 to 60.9% in 2024, a drop to about 858 million people.[R21] The share of the population aged 60 and over was 22% in 2024 and is projected by China's own National Bureau of Statistics to reach 34.9% by 2050. FIG. 09 — China's population, actual and UN medium-variant projection, 2022 to 2100. The 2022, 2025 and 2050 points are as published by the UN Population Division, reported via Pew Research Center, RAND Corporation and worldpopulationclock.net [R21]. *The 2030 point (1.387 billion) is Samso's own calculation, subtracting a separately-published 2025-to-2030 decline estimate from the 2025 total, not a figure the UN published directly for that year. †The 2100 figure varies between published sources (roughly 630 to 640 million); this chart uses the midpoint. The urbanisation story Japan, South Korea and Britain went through in the 20th century relied on the overall population still growing while cities grew (see FIG. 09). China's does not. It's urbanisation running directly against demographic headwind. The closer comparison is what happened inside Japan after 1990, where a handful of major metropolitan areas kept growing and absorbing what population remained, while huge stretches of smaller cities and rural regions hollowed out. China appears to be trying to engineer that outcome deliberately, through the megaregion strategy and the land-allocation link to population flow, rather than let it happen by accident the way it did in Japan. 7.00 — AI ARRIVES AT THE SAME MOMENT Beijing is legislating against the risk it created The third force is the one that connects most directly back to jobs, and it is well documented, not speculative. China's National Bureau of Statistics reported youth unemployment, for people aged 16 to 24 excluding students, at 15.6% in May 2026, down from 16.9% in March, against a national headline rate of 5.1%.[R22] That's roughly three to four times the prime-age rate, and the gap holds up across every age bracket the NBS reports (see FIG. 10). FIG. 10 — China unemployment by age bracket, May 2026. *Excluding students. Source: China National Bureau of Statistics, as reported by hellochinatech.com [R22]. The clearest single line on why AI is central to this picture comes from inside China's own policy establishment, not from an outside critic. Cai Fang, a prominent labour economist, wrote that "job destruction often precedes and outweighs job creation."[R23] He went on to argue that while technological progress ultimately creates new jobs, the speed and scale of AI penetration and automation could produce long-term employment shocks, and called for more investment in human capital and social welfare protection in response. SAMSO TAKE Beijing is not pretending this risk doesn't exist. A Beijing court ruled last year that dismissing employees solely in order to replace them with AI is illegal, and courts in Hangzhou and Beijing have separately sided with laid-off tech workers in AI-related disputes.[R24] That's a government treating this as a live policy risk rather than a distant theoretical one (see FIG. 11). FIG. 11 — The Supreme People's Court in Beijing. Chinese courts have already ruled against employers dismissing staff specifically to replace them with AI. Photo: Pauloleong2002, Wikimedia Commons, 11 July 2023, CC BY-SA 4.0.[R44] SAMSO TAKE Policy proposals under discussion include capping automation-driven layoffs at 30% of a workforce in any single role, and requiring companies to reinvest a share of their automation cost savings into worker retraining, modelled partly on similar measures already in place in Spain, South Korea and Japan. The position this puts Beijing in is worth stating plainly. The government wants China to lead in AI adoption for the same reason every major economy does, because it drives productivity and competitiveness. But rapid AI adoption in exactly the labour-intensive manufacturing, logistics and services sectors that employ hundreds of millions of people is also the most obvious way to make the youth unemployment problem worse, in the same cities the urbanisation plan is trying to fill. Is Beijing's own labour-protection legislation strong enough to hold that tension for another five years, or does the maths eventually force a choice between the AI race and the employment numbers? This isn't only a Chinese preoccupation. Dario Amodei, the co-founder and chief executive of Anthropic, one of the world's leading AI companies, has warned publicly that AI could eliminate as much as half of all entry-level white-collar jobs within one to five years, and has said the world could end up with fast GDP growth alongside high unemployment and rising inequality at the same time (see FIG. 12).[R30] FIG. 12 — Dario Amodei, co-founder and chief executive of Anthropic, meeting with then UK Prime Minister Rishi Sunak in May 2023. Photo: UK Prime Minister's Office, 10 Downing Street, 24 May 2023, CC BY 2.0.[R45] SAMSO TAKE His own proposed responses cover similar ground to Beijing's rules from a different starting point, including a tax on AI-driven automation, progressive taxation of AI companies, universal basic income, and mandatory pre-release testing and auditing of AI models before they reach the public.[R31] When asked what happens if the disruption outruns the economy's ability to adapt, his answer was direct. "This is the outcome we wanna prevent." Congress has bipartisan efforts under way attempting to legislate some of the same guardrails already written into Chinese court rulings and workforce caps. Beijing and Washington are approaching the same problem from very different political systems, using very different tools, and arriving at a similar list of proposed fixes. Support Samso Independent, organic, and free to read We produce independent organic content and need your support to keep that going. Every contribution covers the cost of research, keeps the site running, and buys the time to write the next piece. A$10, A$20, A$30 or an amount you choose One time, weekly, monthly or yearly A donation buys nothing and unlocks nothing. It keeps the work going. Support Samso → 8.00 — WHAT THE FORECASTS ACTUALLY SAY A managed slowdown, with the weak point named It would be easy to stack these three trends up and reach for a dramatic conclusion. No credible forecaster is doing that, and it's worth being specific about what they're saying instead, because "collapse" and "managed slowdown with a named soft spot" are very different stories. The IMF's 2025 Article IV review of China holds its 2026 growth forecast at 4.5%, describing a "deeper-than-expected" property sector contraction as the main domestic risk, and renewed trade tension as the main external one.[R25] It has called on Chinese policymakers to cut industry subsidies from around 4% of GDP to 2%, and to restructure the debt of local government financing vehicles, the off-balance-sheet entities local governments have historically used to fund infrastructure spending. Other forecasters land in a similar range with varying degrees of scepticism about official figures. Finland's Bank of Finland Institute for Emerging Economies expects actual growth closer to 4% in 2026, slowing further to around 3.5% in 2027 and about 3% in 2028.[R26] UBS expects the drag on GDP growth from the property downturn to narrow from 1.5 to 2 percentage points in 2025 to 0.5 to 1 point in 2026, as the pace of decline in sales, construction starts and investment eases.[R27] One route out of this bind is exports, and China has a genuine lead to sell into. Electric vehicles are the clearest case. Chinese brands dominate the global EV supply chain, and by 2026 Chinese-made electric cars had captured record shares of markets from Australia, where China overtook Japan as the largest source of new vehicles, to the United Kingdom, where a Chinese-linked brand became the country's best-selling new car in a single month (see FIG. 13).[R33] FIG. 13 — A BYD e6, one of the Chinese-made electric vehicles now capturing record market shares from Australia to the United Kingdom, photographed in Hong Kong. Photo: User3204, Wikimedia Commons, 19 May 2024, CC BY-SA 4.0.[R42] Two things narrow that route considerably. The first is tariffs. The United States imposes roughly 100% duties on Chinese-made electric vehicles, and the European Union runs a tiered anti-subsidy tariff running from 17.4% to 45.3% depending on the manufacturer.[R34] Chinese carmakers have adapted, localising production inside the EU, exporting plug-in hybrids and vehicle kits that avoid the tariff, and redirecting volume to markets without one, but the two largest, most lucrative consumer markets in the world remain substantially closed to the finished product that would generate the most value. The second is domestic demand, and this is where the wealth gap matters directly. Household consumption made up only 38.8% of China's GDP in 2020, against more than 60% in the United States and the United Kingdom, and Beijing has only just elevated raising household income to a top-level national policy priority, folded into the 15th Five-Year Plan in March 2026, specifically because weak consumer spending is seen as a bottleneck to growth.[R35] China's Gini coefficient, a standard measure of income inequality where zero is perfect equality and 100 is total concentration in one household, exceeds the United Nations' own warning threshold of 40 on every published estimate, official or independent, though the exact figure is disputed. Official statistics put it at 46.5, other estimates put it above 57. Urban incomes run more than double rural incomes on average.[R35] A country that has concentrated as much of its income gains as China has in its wealthier coastal cities cannot simply ask its regional population to spend its way to the growth that exports aren't fully delivering. The people any consumption-led recovery needs to spend more are disproportionately the people who have benefited least from the growth so far. That's a genuine constraint on trading out of the debt problem, not just a footnote to it. SAMSO TAKE Here's where these separate weak points could compound rather than simply add up, and this isn't a Samso hypothesis, it's a documented finding with named researchers behind it. The Stockholm economist Per Krusell, working with Pedro Brinca, Hans Holter and Laurence Malafry, ran fiscal-policy models across a large set of countries and found a strong, direct correlation between wealth inequality and the size of a country's fiscal multiplier, how much extra growth a dollar of government stimulus or spending actually generates.[R36] Separately, the UC Berkeley economist Alan Auerbach, working with Yuriy Gorodnichenko and Daniel Murphy, built a model showing inequality has large negative effects on output partly because it dampens the very fiscal stimulus governments rely on to offset a downturn.[R49] The mechanism both teams point to is the same one. Wealthier households save a larger share of any income they receive, while poorer households would spend nearly all of it if they had it. Apply that to China's specific situation and the real risk is interaction between these weak points, not any single one of them failing in isolation. Urban renewal spending, debt resolution and any future stimulus all depend on that money eventually turning into household income and then into consumption, to generate the follow-on growth policymakers are counting on. If the wealth gap means a large share of the population that would need to do that spending doesn't have the income to do it, the same stimulus dollar does less work in China than it would in a more evenly distributed economy. Nobody has published a number for how much smaller China's multiplier actually is, and this is Samso's own inference from the pieces already on the table, not a forecast. But it's a real, worst-case mechanism worth naming plainly. The various holes in this piece, the debt nobody can fully see, the exports tariffs are blocking, the population that's shrinking, and the wealth gap limiting consumption, might not simply add up to a slower recovery. They could multiply into one. That's a managed deceleration, not a crash scenario, with the property sector named consistently as the identified weak point, not an unknown one. The population decline and the AI-jobs tension aren't yet showing up as separate line items in these growth forecasts. That absence says something. The forecasting community is currently treating this as a property and trade story, not yet as a demographic-and-automation story. Whether that changes over the next year or two is arguably the more interesting open question than any single growth number. 9.00 — WHERE THIS LEAVES THE READER China gets a slower version of Japan's path, or a harder one China is trying to build the urban capacity for a growing, consuming population using a policy playbook designed a decade ago, at the exact moment its population is shrinking and the jobs that would justify the move are increasingly exposed to automation. That is a genuine structural tension, documented in the government's own statistics and in the words of its own economists, not an outside critique imposed on the story. On the evidence currently available, this is an economy attempting a difficult transition with fewer of the tailwinds it had the last time it tried something this size, not one heading for collapse. Whether it plays out closer to Japan's outcome, three lost decades followed by a real recovery, or something slower and harder given the extra weight of a fragmented, opaque debt structure and two pressures Japan never faced at once, is the open question this piece set out to frame rather than settle. Nobody, including the IMF, the Bank of Finland's own researchers, or China's former chief statistician, is claiming to know the answer with confidence, and that uncertainty is itself the thing worth taking seriously before anyone prices in either outcome. SAMSO TAKE For an ASX investor, the practical version of this question isn't "will China collapse," which the evidence doesn't support, it's "how long does this take, and does China actually have a working route back." Japan's own answer took three decades, and it worked because Japan could simply wait, its debt was visible and contained, and the country never needed a specific growth engine to fire in order to recover. China's two obvious growth engines, exporting its way out through electric vehicles and spending its way out through domestic consumption, are both constrained in ways this piece has set out plainly, and the fiscal-multiplier risk covered above means those constraints could reinforce each other rather than stay separate. A Japan-style outcome is still possible, but it's a harder path to bet on without knowing which lever China actually pulls, or whether either one works as well as the plan assumes. Part 2 looks at where the capital tied up in this transition is actually flowing, and which ASX and TSX-listed producers are closest to benefiting from it, tungsten, tin and copper on one side, and iron ore on the other. If the property-led growth model really is giving way to an AI-led one, which side of that trade is actually positioned to benefit first? Iron ore deserves a place in that discussion too, and not for the reason that gets assumed first. The obvious hope for Australian producers is that the global push toward green steel, made using hydrogen-based direct reduction rather than a coal-fired blast furnace, becomes a catalyst for renewed interest in the sector (see FIG. 14). FIG. 14 — The BHP Yandi mine in the Pilbara, Western Australia, seen from the air. Standard Pilbara ore runs 56% to 62% iron, short of the roughly 67% grade direct reduction steelmaking needs. Photo: Calistemon, Wikimedia Commons, 30 June 2022, CC BY-SA 4.0.[R50] SAMSO TAKE The problem is that direct reduction technology needs ore grading above roughly 67% iron, and standard Pilbara ore runs 56% to 62%, five to nine percentage points short.[R46] That's the opposite of a tailwind. Australian iron ore export revenue is forecast to fall from around A$116 billion to roughly A$97 billion by 2026 to 2027, and the grade gap against higher-grade rivals such as Guinea's Simandou project, at around 65% iron, and Brazilian ore is part of why.[R47] Fortescue is the one major Pilbara producer betting seriously on beneficiation and green hydrogen technology to close that gap, having committed to net-zero Scope 3 emissions by 2040, a target Rio Tinto and BHP have not matched.[R48] Whether that bet pays off before the grade problem costs Australia the higher end of the green steel market is a real open question for Part 2, not a settled one. Samso Access Some of the work does not appear here Samso holds back a set of Insights for members. Same method as this piece, the same sourcing and the same counterweights, on companies and sectors that do not make the public feed. Every member-only Samso Insight unlocked Full access to the gated Insights archive New member-only pieces included as they publish Samso News, Coffee with Samso and Research Workups will join the gated set later. A$20 a month · cancel any time See what is inside → THE VOCABULARY, IN PLAIN ENGLISH Hukou China's household registration system. It historically tied a person's access to schooling, healthcare and welfare to wherever they were officially registered, not wherever they actually lived. LGFV (local government financing vehicle) An entity a local Chinese government sets up specifically to borrow money outside the official limits Beijing places on local government debt. Article IV review The IMF's regular annual health check of a member country's economy and policies. Every reference below carries an identifier, [R1] to [R27], matching the marker used at its first point of use in the body text above. Where a specific figure or quotation is checked, the identifier appears again at that point. [R1] ArchDaily / Big Think, on the original 2013 250-million rural-to-urban urbanisation plan. [R2] State Council Information Office (english.scio.gov.cn), on the 15th Five-Year Plan for Urban Regeneration, 2026 to 2030. [R3] South China Morning Post, on the scale and scope of the urban renewal campaign's expected investment. [R4] Bloomberg, on the 19-megaregion plan and Hong Kong's Northern Metropolis project. [R5] South China Morning Post, on hukou reform under the current urbanisation action plan. [R6] Peterson Institute for International Economics, on the two-tier land market and population-linked land supply. [R7] Wikipedia (Chinese real estate crisis, 2020 to present), citing Bloomberg Economics, on real estate's share of China's GDP. [R8] The Washington Times, on the Evergrande founder's sentencing, 20 August 2026. [R9] International Investment, citing Bloomberg, on 2026 property investment, construction starts and sales figures. [R10] Yahoo News / Ghost Cities reporting, on the "three red lines" policy reversal. [R11] Fortune, citing Reuters, on He Keng's remarks as former deputy head of China's National Bureau of Statistics. [R12] Yahoo News / Ghost Cities reporting, citing the Wall Street Journal, on Evergrande's unfinished presold units. [R13] Yahoo News / Ghost Cities reporting, citing the Brookings Podcast, on Kenneth Rogoff's April 2026 remarks. [R14] Wikipedia (National debt of Japan) and EBC Financial Group, citing the IMF and Japan's Ministry of Finance, on Japan's debt-to-GDP ratio and its domestic, Bank of Japan-heavy holder base. [R15] US Congressional Research Service, citing IMF estimates, on China's explicit local government debt and LGFV debt as a share of GDP. [R16] Fortune, on independent analyst estimates of China's total public sector debt burden. [R17] S&P Global Ratings, via Reuters, on the earlier "debt iceberg" characterisation of China's hidden local government debt. [R18] CKGSB Knowledge, on provincial-level LGFV debt concentration, including Guizhou, Inner Mongolia, Ningxia, Liaoning and Qinghai. [R19] Caixin Global, on Beijing's 12 trillion yuan hidden local debt resolution package and the remaining LGFV operational debt. [R20] Washington Post and Pew Research Center, citing UN Population Division data, on China's population peak and decline. [R21] Pew Research Center, RAND Corporation, Washington Post and China Briefing, citing China's National Bureau of Statistics, on population and working-age population figures. [R22] hellochinatech.com, citing China's National Bureau of Statistics' May 2026 release, on youth unemployment by age bracket. [R23] chinatalk.media and allwork.space, on Cai Fang's published views and the wider labour policy debate. [R24] Yahoo Finance / GuruFocus reporting, on court rulings against AI-driven dismissals. [R25] IMF, 2025 Article IV Consultation for China, on the 2026 growth forecast and named risks. [R26] Bank of Finland Institute for Emerging Economies (BOFIT), on the 2026 to 2028 growth outlook. [R27] UBS, China Outlook 2026 to 2027, on the narrowing property drag estimate. [R28] Sparktour, "Skyline of Shenzhen from Hong Kong," Wikimedia Commons, 29 August 2019, CC BY-SA 4.0. Reproduced photograph, not an original Samso illustration. [R29] 中少, "Unfinished buildings in Maoming," Wikimedia Commons, 27 July 2021, CC BY-SA 4.0. Reproduced photograph, not an original Samso illustration. [R30] "Inside the Mind of Anthropic CEO Dario Amodei," The Circuit, Bloomberg Originals, extended interview transcript, 2026, on Dario Amodei's stated estimate of AI-driven entry-level white-collar job losses and the "fast GDP growth, high unemployment" scenario. [R31] "Inside Anthropic, the $965 Billion AI Juggernaut," The Circuit, Bloomberg Originals, interview transcript, 2026, on Dario Amodei's proposed policy responses, including automation taxation, progressive taxation of AI companies, universal basic income and mandatory pre-release testing and auditing. [R32] Next City, "Geographers Studied Shrinking U.S. Cities for Two Years," on rightsizing plans in Youngstown, Detroit, Flint and Buffalo. ScienceDirect, "Shrinking cities, an unfit term for American urban policy," on the German Federal Cultural Foundation's Shrinking Cities Project and the Schrumpfende Städte program in the former East Germany. [R33] GreenCars, "The EVs the Rest of the World Is Already Driving," on Chinese EV market share gains in Australia and the United Kingdom, 2026. [R34] Transport & Environment analysis and the European Parliamentary Research Service, on EU anti-subsidy tariffs on Chinese-made electric vehicles. Rare Earth Exchanges, on the roughly 100% US Section 301 tariff on China-origin EVs. [R35] Caixin Global, "China Targets Income Growth to Rebalance Its Economy," March 2026, on household consumption's share of GDP, the Urban and Rural Residents' Income Growth Plan, China's official Gini coefficient and the urban-rural income ratio. Statista and World Economics, on independent Gini coefficient estimates for China. [R36] Pedro Brinca, Hans A. Holter, Per Krusell and Laurence Malafry, "Fiscal Multipliers in the 21st Century," Journal of Monetary Economics 77 (2016), on the strong correlation their cross-country models found between wealth inequality and the size of the fiscal multiplier. [R37] CEIC, "Japan Urban Land Price Index: Biggest 6 Cities: Residential," citing the Japan Real Estate Institute, on the index level from 1955 to 2017. Housing Japan and Japan Property Research, on Japan's Ministry of Land, Infrastructure, Transport and Tourism official land price survey, released March 2026, reporting the fifth consecutive annual rise in nationwide land prices. [R38] yt_siden, "Bank of Japan Head Office, Tokyo," Wikimedia Commons, 29 August 2011, CC BY-SA 2.0. Reproduced photograph, not an original Samso illustration. [R39] Glabb, "Beipanjiang Highway Bridge, Guizhou, China," Wikimedia Commons, 21 August 2011, CC BY-SA 3.0. Reproduced photograph, not an original Samso illustration. [R40] Dblcut3, "Youngstown skyline as seen from Wean Park," Wikimedia Commons, 6 August 2020, CC BY-SA 4.0. Reproduced photograph, not an original Samso illustration. [R41] Windmemories, "Residential buildings developed by Evergrande in Pingyuan New Area, Yuanyang," Wikimedia Commons, 5 March 2021, CC BY-SA 4.0. Reproduced photograph, not an original Samso illustration. [R42] User3204, "2021 BYD e6 (front), Tai Po, Hong Kong," Wikimedia Commons, 19 May 2024, CC BY-SA 4.0. Reproduced photograph, not an original Samso illustration. [R43] Popolon, "Ordos Museum in Ordos City, Inner Mongolia," Wikimedia Commons, 5 May 2015, CC BY-SA 4.0, architects Ma Yansong, Yosuke Hayano and Dang Qun of MAD Architects. Reproduced photograph, not an original Samso illustration. [R44] Pauloleong2002, "Supreme People's Court in Beijing," Wikimedia Commons, 11 July 2023, CC BY-SA 4.0. Reproduced photograph, not an original Samso illustration. [R45] UK Prime Minister's Office, 10 Downing Street, "Dario Amodei in 2023," Wikimedia Commons, 24 May 2023, CC BY 2.0. Reproduced photograph, not an original Samso illustration. [R46] IEEFA, "Australia faces growing green iron competition from overseas," and CSIRO, "Australia's green metals gambit," on direct reduction ore-grade requirements above 67% iron against standard Pilbara grades of 56% to 62%. [R47] Discovery Alert, "BHP's Green Steel Shift: The Pilbara's Grade Challenge," on the forecast decline in Australian iron ore export revenue to 2026 to 2027 and the competitive position of Simandou and Brazilian ore. [R48] IEEFA, "Opportunities and challenges for the Pilbara amid the accelerating steel technology transition," on Fortescue's HYFOR and smelter reduction technology partnerships and its 2040 net-zero Scope 3 commitment, compared with Rio Tinto and BHP. [R49] Alan J. Auerbach, Yuriy Gorodnichenko and Daniel Murphy, "Inequality, Fiscal Policy and COVID19 Restrictions in a Demand-Determined Economy," European Economic Review 137 (2021), on inequality's negative effects on output and on the effectiveness of fiscal stimulus. [R50] Calistemon, "BHP Yandi mine seen from the air, June 2022," Wikimedia Commons, 30 June 2022, CC BY-SA 4.0. Reproduced photograph, not an original Samso illustration. Cover photo: Shanghai Skyline 2012, Wikimedia Commons user Yourusernamewillbepublic2. CC0, via Wikimedia Commons (https://commons.wikimedia.org/wiki/File:Shanghai_Skyline_2012.jpg). Reproduced photograph, shown faded under a navy overlay on the article thumbnail. Open item before this is considered final. The youth unemployment figures in circulation for 2026 range from 15.6% to as high as 18.9% depending on source. This piece uses 15.6%, traced directly to an NBS data release, and treats any higher figure as unverified. The Samso Way – Seek the Research Our mission is simple: cut through the noise and spotlight what matters, genuine stories, grounded insights, and real opportunity. Investors can explore our four core platforms: Coffee with Samso Samso Insights Samso News Samso Research There may be numerous paths to success in investing, but the common thread among successful individuals is that they remain committed to making informed decisions. Equip yourself with the right knowledge and tools, and you will be well on your way to achieving your financial goals. Most importantly, investors need to be absolutely diligent in understanding their own risk-reward tolerance and capabilities. Never bite off more than you can chew. As they say, Rome wasn't built in a day, and the Great Wall stood because it took centuries to complete. The Samso Philosophy: Stay curious. Stay sharp. And remember, digging deeper always uncovers the real value. In Life, there is no such thing as a Free Lunch. Never bite off more than you can chew is my parting comment. Happy Investing, and the only four-letter word you need to know is DYOR. To support our independent work, please head over to our Support Page and give us a helping hand in any of the ways listed. This is a new initiative for the Samso Platform, and it was always the concept of Samso when we started this journey in 2018. Disclaimer This article is general information only. It is not financial product advice and it does not take into account any reader's objectives, financial situation or needs. It contains no price targets, no valuations and no recommendation to buy, hold or sell any security. Companies named in this article are used as historical illustrations of events that have already occurred and no view is expressed on their current merits. Readers should consider obtaining advice from a licensed financial adviser and should conduct their own research before making any investment decision. Samso and its associates may hold interests in companies mentioned. Read full Disclaimer. About Samso Samso is a trusted platform that equips dedicated investors with up-to-date industry knowledge and insights from top CEOs and thought leaders. By staying informed on business advancements and market trends, investors can enhance their financial decisions through a combination of expert guidance and their own research. The Investor's Playbook is an open-ended, evergreen educational series about how to think about investing rather than what to buy. Samso Insights | www.samso.com.au | An Investor Lens on ASX-Listed Companies

Part 1 measured the rock at Bengwenyama and found it good. Part 2 looks at the company that owns it, and at three things the market has not put a number on. A second product the published economics do not contain. A discount two research houses apply and neither of them explains. And a commodity whose best deposits are so concentrated in one country that investors outside it have almost nothing to compare this with. Samso Insights Platinum Group Elements ASX Small Caps Investor Awareness WHAT THIS PIECE COVERS 1.00 WHERE PART 1 LEFT OFF The three tests Part 1 ran, the five this piece runs, and why there is so little to compare this with, on ground, grade and power. 2.00 WHO ACTUALLY OWNS BENGWENYAMA The corporate chain, the community's 30 per cent, the loan behind it and the decade of litigation that kept this ground free. 3.00 TEST FOUR, A CONVENTIONAL FLOWSHEET What the July 2026 metallurgical work changed, and the second product that is in none of the published numbers. 4.00 TEST FIVE, A CONCENTRATE SOMEONE WILL BUY Chrome as a penalty and as a product, and why an unsigned offtake is a step not yet reached. 5.00 TEST SIX, A RESERVE RATHER THAN A RESOURCE A declared reserve of 6.29 million ounces, the three qualifications on it, and why it is more likely a floor. 6.00 TEST SEVEN, PERMITS A mining right granted after nearly three years, and the licences and steps still outstanding. 7.00 TEST EIGHT, BEING ABLE TO RAISE THE MONEY About 1.7 times the market capitalisation, measured against a mine the company no longer intends to build. 8.00 WHAT THE RESEARCH SAYS, AND WHO PAID FOR IT Who covers this company, who pays them, and the discount neither of them itemises. 9.00 THE VERDICT Six tests answered, two open, and the three things that have no line on the scorecard at all. 10.00 FROM WHERE I STAND Noel's own commentary. A note on the vocabulary and the sources. Technical and mining terms are explained in plain English where they first appear, and all of them are collected in one box at the back. Every source carries an identifier, [R1] to [R40], and that identifier appears in the body text at the point the source is used, so any figure can be traced to where it came from. Where this article builds a model from released data rather than repeating a company statement, it says so at that point and labels it as Samso’s own reasoning. A model is not a prediction. 1.00 — WHERE PART 1 LEFT OFF The eight tests, and how far Part 1 got Part 1 set out eight tests a platinum group element project has to pass before it becomes a mine. It ran the first three, which are grade, the orebody, and what the plant costs to build. PLATE 01 is what the ground looks like now. A camp, a core yard, a road and a fence. Everything in this piece that has not happened yet would be built on that flat ground in the middle of the picture. PLATE 01 The exploration camp and core yard at Bengwenyama, with community housing and the Eastern Limb hills behind. Photograph from the Southern Palladium investor presentation of 2 September 2026 [R36]. Samso has not visited the site and cannot date the photograph. Bengwenyama passed on grade. The company reports 9.78 grams per tonne on a 7E basis, and a 1982 study sampling the UG-2 across the whole Bushveld put the central eastern sector at 10.05 parts per million against a Bushveld-wide average of 6.93. On the orebody, the reef is narrow at about 73 centimetres, a dome has removed the reef from part of the property, and potholes take about 17 per cent of the reef area. On what it costs to build, Stage 1 works out at roughly US$1,100 of capital for every ounce of annual production capacity, against roughly US$3,900 for Chalice Mining's Gonneville on the same calculation. Those three tests measure the rock, so the answers to them do not change from one year to the next. The five remaining tests are about the company. Whether the processing flowsheet is conventional. Whether anyone will buy the concentrate. Whether the resource has been converted into a reserve that stands up. Whether the permits are in hand. And whether a company of this size can raise US$279 million. Most of those now have an answer. What this piece is really about is the part that does not, because that is where the value in Bengwenyama has yet to be priced. A second product the published economics do not contain. A discount two research houses apply to the project and neither of them explains. And a commodity so concentrated in one country that an investor outside it has almost nothing to measure this against. Part 1 did not ask who owns the project. Southern Palladium holds 70 per cent of it, and what that 70 per cent entitles the company to do changes how three of the five remaining tests read. Section 2.00 sets out the ownership before the tests are run. Why there is so little to compare Bengwenyama with Platinum group metals are not a commodity most investors outside South Africa have had to learn, and there is a reason for that. The United States Geological Survey states it in one line, that "the largest resources and reserves are in the Bushveld Complex in South Africa" [R32]. The same table puts South Africa's platinum group metal reserves at 63,000,000 kilograms against a world total of more than 76,000,000 [R32]. So roughly four fifths of the world's reserves are in one country, and most of that is in one geological complex. Bengwenyama is in it. Russia is the other name on that table, and it is there for production rather than reserves. The USGS calls Russia "the world's leading producer of mined palladium" [R32]. In 2024 Russia mined 89,000 kilograms of palladium against South Africa's 82,600. On platinum the position reverses hard, with South Africa at 126,000 kilograms against Russia's 22,000. The world's palladium supply leans on Russia and the world's resource base is South African, and neither country appears on anybody's list of easy places to build a mine. That is why the comparison set is thin. The nearest ASX-listed comparison in this piece is Chalice Mining's Gonneville, which needs about US$3,900 of capital for every annual ounce against Bengwenyama's US$1,100. An Australian investor reaches for Gonneville because there is not much else on the board to reach for. FIG. 01 makes the same point from the other direction. On the ASX there is almost nothing to measure Bengwenyama against. On the ground it is surrounded by mines that have been running for years, which is why the project keeps being described as being in an established district rather than a frontier one. FIG. 01 Bengwenyama on the Eastern Limb of the Bushveld Complex, with the operating mines the company names as neighbours. Reproduced from the Southern Palladium investor presentation of 2 September 2026, slide 9 [R36]. The map is the company's own. Samso has not checked the positions marked on it against a survey source. Power belongs in that comparison, and it is one of the places where South Africa is plainly ahead. Business electricity in South Africa averaged about US$0.105 a kilowatt hour across 2023 to 2026, against about US$0.242 in Australia and a world average of about US$0.164 [R35]. On those figures a South African operation pays roughly 40 per cent of what an Australian one pays for the same unit. For a mine, where power is one of the largest running costs over a thirty year life, that is not a rounding difference. Two things have to be read against it. The direction of travel is up. Eskom's approved standard tariff increase for 2026 and 2027 is 8.76 per cent from 1 April 2026 [R34], industrial prices have roughly doubled over the past decade, and large users have been moving to self generation to get away from the increases [R33]. And the comparison above comes from a commercial price aggregator rather than a regulator, and its business category is not the specific tariff a mine on a 55 MVA connection would negotiate. Bengwenyama holds a budget quotation for that connection and nothing signed, so its own delivered cost is not established. F-21 records that gap. On reliability the change is more recent and more striking. South Africa went more than a year without load shedding to August 2026, and Eskom's energy availability factor reached about 82 per cent on 24 July 2026 [R33]. Load shedding was the loudest single argument against building anything in South Africa for most of the past decade. It has not been a live constraint for a year, and that is a genuine change in the case for building here rather than a forecast that it will continue. Power is one part of a wider change in South Africa. Most of it is recent. Eskom forecast a surplus of about 6 gigawatts of peak capacity over the 2026 winter, and said no load shedding was expected even if unplanned outages reached 14 gigawatts [R37]. The winter then ran through without load shedding. Transnet concluded network access agreements with eleven private train operating companies in May 2026, together bringing about 24 million tonnes of freight capacity [R38]. It is the first time private operators have run trains on the national freight network. That matters more to Bengwenyama than it would to most mines. The chrome stream in section 3.00 is a bulk product measured in hundreds of thousands of tonnes a year, and bulk products move by rail or they do not move economically. S&P Global Ratings upgraded South Africa in November 2025, its first upgrade of the country in about twenty one years. The foreign currency rating went to BB and the local currency rating to BB+, both with a positive outlook, and S&P affirmed both in May 2026 [R39]. BB+ is one notch below investment grade and BB is two. Fitch upgraded South Africa in June 2026. South Africa was removed from the Financial Action Task Force grey list on 24 October 2025 [R40]. The grey list is the task force's register of countries under increased monitoring for money laundering and terrorist financing controls. Sitting on it raises the cost and the paperwork of moving money into and out of a country, which matters to a project that has to import plant and export concentrate. Samso reads those four together rather than one at a time. Business and government have been working the same problems, some markets have been opened to private operators, and mining reform is on the agenda. That is a reading of direction and it is Samso's, not a measured fact. None of it makes South Africa a Tier 1 jurisdiction and this piece does not argue that it does. The Mining Charter, section 11 of the Act and the permitting queue are all still what they were, and each is dealt with later. What has changed is the state of the two things investors named most often as reasons not to build here. The power system carried a winter without cutting anyone off, and the country is no longer under increased international monitoring for financial crime controls. The jurisdiction is also why grade and volume carry more weight here than those words usually suggest. A deposit is economic when it can carry the cost of being built and then survive the price cycle once it is running. Grade and volume are what allow that, because a high grade at scale keeps the cost of each ounce low enough to trade through the bad years rather than stopping. Bengwenyama's numbers for that test are already public. Nine point seven eight grams per tonne on a 7E basis, a 6.29 million ounce Probable reserve inside a 40.25 million ounce resource, and about US$1,100 of capital for every annual ounce of capacity. SAMSO TAKE The reconciliation this piece asks a reader to make. The same single fact produces both halves of the argument. The best platinum group element ground is in South Africa, which is why Bengwenyama has the grade and the volume, and it is also why it carries the jurisdiction risk. You do not get the first without accepting the second. An investor who rules out South Africa has not moved to safer ground, they have moved to a smaller opportunity set whose other end is Russia. This is Samso's view rather than a finding, and Section 10.00 carries the fuller argument in Noel's own words. It is a reason to weigh the discount rather than accept it. It is not a reason to buy anything. 2.00 — WHO ACTUALLY OWNS BENGWENYAMA The ownership chain, from the ASX listing down to the farms Southern Palladium Limited is an Australian company listed on the ASX under SPD and on the Johannesburg Stock Exchange under SDL. The chain from the listed parent down to the two farms runs through four companies. The announcement most investors read compresses it to a single line. The chain runs like this. Southern Palladium Limited owns 100 per cent of SPD (RSA Holding) Pty Limited in Australia, which owns 100 per cent of SPD South Africa Proprietary Limited, which owns 70 per cent of Miracle Upon Miracle Investments (Pty) Ltd, a South African company known in every company document as MUM. MUM holds the project 100 per cent [R1] [R2]. The two Australian and South African intermediate companies were incorporated on 12 May 2022, and SPD South Africa acquired the 70 per cent of MUM on 26 May 2022 [R1]. The mining right announcement of 10 August 2026 does not name either of them. It says only that "the Project is 100% held by Miracle Upon Miracle Investments (Pty) Ltd (MUM), in which Southern Palladium holds a 70% ownership interest" [R3]. FIG. 02 is the company's own diagram of the structure, published for the listing in June 2022. The chain below Southern Palladium Limited is the same one the company operates today. What has changed is the top of the chart, where the shareholdings have moved with four years of capital raisings. FIG. 02 Southern Palladium’s own corporate structure diagram, drawn for the listing. Reproduced from the pre-quotation disclosure of 3 June 2022, section 4(b), released to the ASX and JSE on 6 June 2022. This is the structure on completion of the acquisition of MUM and of the initial public offering. The chain below Southern Palladium Limited is unchanged today. The shareholdings above it are not. Nurinox held 12.3 per cent of the listed company at listing and 5.7 per cent at 10 August 2026. “Propriertary” is the company’s own spelling. TABLE 01 sets the same chain out with the jurisdictions and the registration numbers, and brings the holdings to August 2026. What the 70 per cent entitles Southern Palladium to do Southern Palladium sets out the terms of the joint venture in Note 6 of its 2025 Annual Report, in these words [R2]. Most coverage of the company leaves this out. "Whilst the company owns 70% of the voting rights, the terms of the arrangement require unanimous consent over decisions relating to relevant activities of MUM. The Company has determined that the relevant activities for its joint arrangement relate to the operating and capital decisions of the arrangement, such as the approval of the expenditure programme for each year." Because of that, Southern Palladium does not consolidate MUM into its accounts. It equity-accounts MUM as a joint venture. In plain terms, operating and capital decisions at MUM need both partners to agree, so the 30 per cent holder can block the annual budget. Southern Palladium takes 70 per cent of the economics without the ability to decide anything alone. Its own accounts describe the arrangement as a 70 per cent owned joint venture, and that is the accurate description. The lack of control shows up in every quarterly report. The cash figure Southern Palladium publishes carries the rider that it "excludes cash held by the Company's 70% subsidiary, Miracle Upon Miracle Investments (Pty) Limited" [R4]. At 30 June 2026 that was A$23.879 million at the Southern Palladium level and A$0.98 million at MUM, reported separately rather than added together. Who holds the other 30 per cent The 30 per cent belongs to the Bengwenyama-ya-Maswazi community, the people living on and around the two farms. At the time of the 2022 listing it was held by Nurinox Investments Proprietary Limited, a company wholly owned by that community [R5]. The mining right announcement of August 2026 describes the holder as "the Bengwenyama Traditional Council on behalf of the Bengwenyama Community" [R3]. Whether the register at MUM changed between 2022 and 2026, or whether the later wording is simply looser, is not established in any document Samso could obtain. The community also holds shares in the listed parent. Nurinox held 7.55 per cent of Southern Palladium at 14 October 2025 [R2] and 5.7 per cent at 10 August 2026 [R3], the fall being dilution from share issues rather than any sale. The community therefore holds an interest in the project twice over, once through MUM and once through the listed parent. The 2025 Annual Report states the total in these words [R2]. "The Community holds an effective economic interest in the Company of 35.3% via a 30% direct holding and a further indirect holding of 5.3% through its 7.55% shareholding in SPD." On the August 2026 shareholding of 5.7 per cent, the same arithmetic gives about 34 per cent. That calculation is Samso's, using the company's own method and its own updated figure. No company document states the updated total. PLATE 02 is company material, and it is here as company material rather than as evidence. Social investment and a shareholding are separate things, and only one of them appears in the accounts. PLATE 02 Young players at a Southern Palladium sponsored community sports event near Bengwenyama. Photograph from the Southern Palladium investor presentation of 2 September 2026 [R36]. Samso cannot date the photograph or identify the event, and reproduces it as company material rather than as verification of any community programme. How the community's 30 per cent is funded Broker commentary and investor conversation both tend to describe a community holding of this kind as a free carry, meaning the partner takes its share of the profits without paying its share of the costs. No document Samso obtained uses that phrase about Nurinox or the Traditional Council. The documented arrangement is a loan. Southern Palladium funds MUM through a loan. The 2025 Annual Report describes a loan facility and services agreement, originally dated about 20 May 2021 and amended in March 2022, under which Southern Palladium grants MUM a prospecting facility of A$19 million for prospecting operations [R2]. At 30 June 2025 the balance was A$12,868,401 of principal plus A$1,792,937 of accrued interest, a total of A$14,661,338, and the whole amount was written down to nil because the company judged that it is uncertain whether MUM can repay it at this stage of the project. The loan carries interest at the United States bank prime rate, 7.50 per cent at 30 June 2025, not compounded. It is repayable in full no later than five years after MUM begins commercial mining operations, and Southern Palladium cannot demand it back before then. So the 30 per cent partner has contributed no cash and has not been diluted, which is the economic effect of a carry. The money Southern Palladium has put in is still a debt of MUM. It ranks ahead of any distribution to either shareholder and it accrues interest until it is repaid. This is a loan carry. In the first years of production MUM's cash goes to repaying the loan, which a free carry would not require. SAMSO TAKE Ownership arithmetic an investor can do at home. A project NPV is always a 100 per cent number. Southern Palladium's share of the US$857 million net present value in the Optimised Pre-Feasibility Study is 70 per cent of it, before the shareholder loan is repaid and before any equity the company still has to issue to build the mine. Section 7.00 puts numbers on the second of those. Samso raises this as a question and not as a conclusion. Does the market price the project, or Southern Palladium's 70 per cent share of it? Why this ground was still available in 2022 Bengwenyama is one of the last shallow undeveloped sections of the UG2 reef on the Eastern Limb, in a district that has been mined for close to a century. The reason it stayed undeveloped is legal rather than geological. The ground spent roughly a decade in litigation, which the community won. In Bengwenyama Minerals (Pty) Ltd and Others v Genorah Resources (Pty) Ltd and Others, decided by the Constitutional Court of South Africa on 30 November 2010, the Court set aside the Department's decision to grant prospecting rights over Nooitverwacht 324 KT and Eerstegeluk 327 KT to Genorah Resources [R6]. The applicants were Bengwenyama Minerals, the Bengwenyama-Ye-Maswazi Tribal Council and the trustees of the Bengwenyama-Ye-Maswati Trust. The Court held that the Department should have told the community about Genorah's application and about its effect on the community's own preferent right to prospect on its own land, and should have allowed the community to bring its own application before deciding Genorah's. On consultation, the Court held that an applicant must inform the landowner in sufficient detail about the proposed operations and consult with a view to reaching an agreement. Costs were awarded against the respondents across all courts. A prospecting right was awarded to MUM in June 2015, registered for five years in 2017 and renewed for a further three years in February 2021 [R7]. Southern Palladium then negotiated to acquire 70 per cent of MUM, which is what the June 2022 initial public offering funded. The full ownership history, including the 1940s displacement of the community from the two farms and the successful land claim in 1996, comes from Bridge Street Capital Partners' August 2026 report and Samso has not independently verified it [R7]. Bridge Street are corporate advisers to Southern Palladium. Their disclosures are set out in section 8.00. How Southern Palladium bought its 70 per cent Southern Palladium bought its 70 per cent up front with shares, under an instrument called the Second Amended and Restated Share Exchange Agreement, and put half the consideration at risk against two project milestones. The more common structure, earning an interest by spending money over time, was not used here. The company issued 45,500,000 fully paid ordinary shares to the vendors, valued at the initial public offering price of A$0.50, so A$22,750,000 in total, with no cash paid [R2]. Half of those shares, 22,750,000 of them, were "Clawback Shares" that Southern Palladium could buy back for a nominal A$50 and cancel if two project milestones were not achieved within four years of the offer [R1]. The two milestones, in the company's own words [R1], were "at least 2 million oz PGE (4E) in JORC classified reserves" and "the acceptance by the Regional Manager (as contemplated in the MPRDA) of the formal application by the Company for a Mining Right over the Project". Both were met. The mining right application was lodged on 29 September 2023 and accepted on 17 October 2023 [R4]. The maiden ore reserve of 6.29 million ounces on a 6E basis was declared on 23 October 2024 [R8], which is well above 2 million ounces on any counting basis. The 2025 accounts record the outcome plainly, that "in the current financial year, the clawback condition dropped away as the performance conditions were satisfied" [R2]. The vendors kept all 45.5 million shares. Those shares are the origin of two of the three substantial holdings on the register today. Black economic empowerment, and what the law requires In South Africa the state holds mineral resources as custodian, and a company obtains the right to mine by applying for a grant under the Mineral and Petroleum Resources Development Act of 2002. Section 23(1) of that Act sets the criteria for granting a mining right, and section 23(1)(h) requires the Minister to be satisfied that the grant will expand opportunities for historically disadvantaged persons, in accordance with the charter contemplated in section 100 and the prescribed social and labour plan [R9]. The Mining Charter of 2018 sets a minimum 30 per cent black economic empowerment shareholding for a new mining right, split as 5 per cent to qualifying employees as a non-transferable carried interest, 5 per cent to host communities on the same basis, and 20 per cent to black economic empowerment entrepreneurs [R10]. Most commentary stops at that 30 per cent figure. A High Court decision in 2021 changed its legal status. On 21 September 2021, a full bench of the Gauteng Division of the High Court in Pretoria decided Minerals Council of South Africa v Minister of Mineral Resources and Energy and Thirteen Others [R11]. The Court held that "the 2018 Charter is not binding subordinate legislation but an instrument of policy". The Court reasoned that section 100(2)(a) of the Act does not empower the Minister to make law. It then reviewed and set aside a long list of clauses, most of them directed at existing rights, renewals and transfers, including the re-empowerment obligations and the procurement and enterprise development targets. The Minister did not appeal, and said so on 23 November 2021 [R12]. The 30 per cent requirement for a new mining right application was not struck out. It is now a policy standard rather than a legal duty, which means the Minister may apply it when exercising discretion under section 23(1)(h) and is not obliged to. This is important for an investor to understand. A policy standard can be applied differently to different applicants, and it can change without legislation. Southern Palladium's structure holds 30 per cent of the mining right holder for the community, and the community holds a further 5.7 per cent of the listed parent. The company has prepared a social and labour plan, which the 2025 Annual Report says was developed in compliance with the Act, the 2004 regulations and the 2018 Mining Charter, and has been reviewed by the Department [R2]. No public document explains how a single 30 per cent community block was assessed against the Charter's split of 5 per cent employees, 5 per cent community and 20 per cent entrepreneurs. None says whether an employee scheme exists either. Samso could not find anything addressing either question. F-3 records that gap. On empowerment compliance, the grant itself is the best evidence available. The Department granted the right on 7 August 2026, which means it was satisfied on section 23(1)(h). That is an inference from the decision rather than a published finding, and it still carries more weight than an analyst's view, because the Department had to be satisfied before it signed. Bridge Street's August 2026 report states that "we therefore are of the view that the Bengwenyama project has achieved full BEE status with no further adjustment to ownership required" [R7]. Southern Palladium has never said that in an announcement, and no departmental document Samso saw says it. Bridge Street are corporate advisers to the company, were lead manager to its listing and to subsequent raisings, and the author holds shares in it. Section 11, and what it does in a takeover Section 11(1) of the Mineral and Petroleum Resources Development Act says that a mining right, or an interest in one, or a controlling interest in a company or close corporation, may not be transferred or otherwise disposed of without the written consent of the Minister, "except in the case of change of controlling interest in listed companies" [R13]. That works out differently at different levels of this structure. MUM is unlisted, so any change in the 70 per cent or the 30 per cent would need ministerial consent, and so would any transfer of the mining right itself. Southern Palladium Limited is listed on two exchanges, so a change of control at the listed parent falls within the express exception. On the face of the section, a takeover of Southern Palladium would not need the Minister's consent. A sale of Southern Palladium's stake in MUM would need it. That reading is Samso's, from the wording of the section. No Southern Palladium announcement Samso read mentions section 11 at all. A draft Mineral Resources Development Bill published on 20 May 2025 would have widened section 11 to cover any interest in an unlisted company, and would have made a transfer without consent void [R14]. A correction issued on 9 June 2025 removed the listed-company limb. Samso could not establish the Bill's status as at August 2026, so nothing here should be read as current law. 3.00 — TEST FOUR, A CONVENTIONAL FLOWSHEET The definitive feasibility study metallurgy of July 2026 New processing technology adds risk to both the schedule and the budget. Part 1 used Chalice Mining as the illustration, four years and about A$15 million of metallurgical test work to settle a flowsheet on a project not yet built. Southern Palladium released the results of its definitive feasibility study metallurgical test work on 9 July 2026 [R15]. The numbers moved a long way, and the chromite recovery moved furthest. TABLE 02 sets the results against what the Pre-Feasibility Study had assumed. Chromite recovery went from 30 per cent to 85.6 per cent, which changes what sort of mine Bengwenyama would be. At full Stage 2 production the company now expects "over 400 koz of PGMs and 1 million tons of high-grade chrome concentrate per annum" [R3]. TABLE 02 above is Samso's own compilation from the announcement. FIG. 03 is the company's version of the same comparison, published seven weeks later, and the two agree on every line. FIG. 03 The company's own comparison of definitive feasibility study metallurgy against the Pre-Feasibility Study. Reproduced from the Southern Palladium investor presentation of 2 September 2026, slide 14 [R36]. The underlying test work is the announcement of 9 July 2026 [R15]. The June 2026 quarterly puts the change in these words [R4]. "Historically, chrome revenues were treated primarily as a by-product credit against PGM production costs. The latest DFS results suggest chrome should instead be regarded as an important standalone revenue stream." The published economics still use the old recovery figure and an old chrome price. The Optimised Pre-Feasibility Study of July 2025 assumes chrome concentrate at US$225 a tonne, real and flat for every year of the mine life, and it assumes 30 per cent chromite recovery [R16]. The company's own July 2026 announcement records South African 42 per cent chrome concentrate trading at US$320 a tonne on a cost and freight basis in May 2026 [R15]. So the only published study for Bengwenyama understates chrome on price as well as on recovery. How much that is worth cannot be known until the definitive feasibility study restates the economics. On chrome, the July 2025 numbers are too low on both price and recovery. The flowsheet itself It is an MF2 circuit, which means the ore is milled and floated, then the tails are milled again and floated again. That configuration has been used on UG2 ore in South Africa for decades. Three additions have been made to it. Dense media separation at the front rejects 24 to 31 per cent of the run-of-mine mass at a 3.2 grams per cubic centimetre cut point, lifting the feed grade from 5.2 to 7.2 grams per tonne 3E while losing 1.2 to 2.2 per cent of the platinum group metal. Spirals on the classifier overflow take out chromite between stages. A second chromite flotation stage treats the secondary rougher tails [R15]. Dense media separation separates rock by density in a fluid heavier than water, so the lighter waste floats off and the denser ore-bearing material sinks. It is standard equipment in coal, diamonds and chrome. FIG. 04 is the circuit as the company draws it. The two boxes on the right are the whole argument of this section, because one circuit produces two things somebody could buy. FIG. 04 The MF2 flowsheet, from dense media separation through staged flotation to two saleable concentrates. Reproduced from the Southern Palladium investor presentation of 2 September 2026, slide 12 [R36]. The company states its own position on technology risk in these words [R15]. "No novel technology is being applied, but rather existing technologies are being optimised for the Bengwenyama ore body. Several companies in South Africa have had to retrofit existing plant with processes that enable better chrome recoveries upfront and more efficient downstream MF2 processes." The flowsheet is conventional, which answers the test Part 1 set. The mining method sits in the same place. Bengwenyama is a narrow tabular reef, which means the ore is a thin sheet running for kilometres rather than a thick body a mine can attack from several directions. Following a sheet like that is what South Africa calls conventional mining. Crews drill and blast along the reef and take out as little waste rock as they can. That is routine work in this country. The UG2 is the reef most Bushveld mines are working, and Samso's estimate is that something like two thirds of South Africa's platinum group metal production now comes off conventionally mined UG2. Samso could not verify that share against a primary source, so it is an estimate rather than a measured figure, and F-22 records the gap. What does not depend on the estimate is that the crews, the equipment and the contractors for narrow reef mining are already in the country and already trained. That is not true of every jurisdiction a new platinum group metal project could sit in. Three questions the announcement leaves open The flowsheet is not final. The June 2026 quarterly, released three weeks after the metallurgical announcement, says part of the reason for the definitive feasibility study taking longer is "the assessment of the potential incorporation of Jameson Cells technology into the processing flowsheet" [R4]. Jameson Cells are a proprietary flotation device, commercially proven and widely used, so the technology itself carries little risk. The circuit to be built has not been settled. The recoveries are laboratory results. The announcement is explicit that open circuit test work was used as the input, and that the closed circuit performance was estimated by constructing a grade recovery response curve from that open circuit data [R15]. There is no pilot plant result and no continuous circuit result. Estimated full-plant recoveries built from bench tests on one bulk sample are a normal stage of a feasibility study. They are still an estimate rather than a demonstrated result. The concentrate does not yet meet the chrome specification. The primary PGM concentrate assayed 3.2 per cent Cr2O3 and the secondary 3.9 per cent, against a target the company itself states as under 2.5 per cent [R15]. The company's answer is that "these grades can be lowered in the commercial application, by improving selectivity in the cleaning circuits". That is an intention rather than a result. Section 4.00 explains why the chrome content of the PGM concentrate matters commercially. The completion date has moved three times since January 2026. In January 2026 the definitive feasibility study was expected in the third quarter of 2026 [R31]. On 9 July it was the fourth quarter of calendar 2026 [R15]. On 30 July it was the beginning of 2027 [R4]. On 10 August it was the first quarter of 2027 [R3]. The reasons given for the most recent moves are the work to capture chrome value in the plant design, and 29 days of drilling lost to heavy rain in Limpopo in the first half of 2026. One more thing about the chrome, and it is the largest number in this piece that no released set of economics carries. None of the chrome improvement is in any released set of economics. The US$857 million net present value, the US$279 million peak funding requirement and the 6.29 million ounce ore reserve were all struck on 30 per cent chromite recovery and a US$225 chrome price. The company has told the market that chrome is now a standalone revenue stream, and that it delayed the definitive feasibility study by a quarter to put chrome into the plant design [R3] [R4]. It has not yet put it into the numbers. Whether that is worth a little or a lot cannot be said until the study lands, and saying otherwise would be restating economics without a study behind them, which is exactly what the ASX stopped the company doing in January 2026. FIG. 05 is as close as the company has come to sizing it. About 500,000 tonnes a year of chrome concentrate in Stage 1 and about 1 million tonnes a year in Stage 2, at 42.2 per cent Cr2O3, beside a figure of US$320 a tonne. Hold that US$320 against the US$225 the published economics assume. Multiply the tonnages out and the numbers are large. They are also not revenue in anyone's model yet, and the US$320 is labelled on the slide as an indicative market reference, which means a market price the company is pointing at rather than a price it has been offered. FIG. 05 The company's own sizing of the chrome stream after the July 2026 test work. Reproduced from the Southern Palladium investor presentation of 2 September 2026, slide 15 [R36]. The US$320 a tonne is the slide's own indicative market reference and is not a contracted price. No published Southern Palladium valuation carries chrome revenue at this recovery. The tonnages on FIG. 05 can be run out, and it is worth doing because no released model carries the answer. The published economics were struck on 30 per cent chromite recovery. The definitive feasibility study test work returned 85.6 per cent. If Stage 2 produces about 1 million tonnes a year of chrome concentrate on the new recovery, the old recovery would have produced roughly 350,000 tonnes of it. The extra is about 650,000 tonnes a year. At the US$225 a tonne the published economics assume, that is about US$146 million a year of revenue nobody has modelled. At the US$320 a tonne shown on FIG. 05, it is about US$208 million. Those are Samso's numbers on Samso's assumptions and they are not the company's. They assume the Stage 2 tonnage on the slide holds, that recovery scales the way the ratio suggests, and that a price holds. Each of those can move. The reason for running the arithmetic is not the answer. It is the size of the gap between what the released economics carry and what the plant is now being designed to produce. On the same arithmetic chrome would run at something like a fifth of project revenue rather than a credit against costs, which is what the company's own change of language in the June 2026 quarterly describes. Samso's working assumption is that at this grade and at the all-in sustaining costs the company now publishes, US$969 an ounce in Stage 1 and US$821 in Stage 2, Bengwenyama would sit in the lower part of the global cost curve before any chrome revenue is counted at all. That is an assumption drawn from the published cost figures and not from a cost curve Samso has built or seen. F-23 records it. Verdict on test four. Passed, with the flowsheet not yet fixed. The technology is conventional and the company says so expressly. The circuit is still being changed, and the recoveries behind the economics have not been demonstrated at scale. SAMSO ACCESS Some of the work does not appear here Samso holds back a set of Insights for members. Same method as this piece, the same sourcing and the same counterweights, on companies and sectors that do not make the public feed. Every member-only Samso Insight unlocked Full access to the gated Insights archive New member-only pieces included as they publish Samso News and Research Workups will join the gated set, with Research Workups also available on their own tier. A$20 a month · cancel any time See what is inside → 4.00 — TEST FIVE, A CONCENTRATE SOMEONE WILL BUY The offtake position, and what the company has said about it Part 1 said this test stops more projects than any of the other seven, and that it gets talked about least. Southern Palladium has no offtake agreement for its platinum group metal concentrate. It has no memorandum of understanding, no letter of intent and no term sheet. It has none of those for its chrome concentrate either. Samso checked the company's complete announcement record, 239 announcements from its listing on 6 June 2022 to 14 August 2026. No agreement of any of those kinds has been lodged. The company states this itself. The JORC Table 1 attached to the Pre-Feasibility Study of 28 October 2024 says, under the heading "Marketing arrangements" [R8]. "There are no legal or marketing agreements in place for the Project. Indicative talks have been held with smelters with expression of interest indicated for offtake of concentrates." The same table, under customer analysis, adds that "no customer and competitor analysis has been undertaken". The company has talks, and it has described them the same way for two years. The Optimised Pre-Feasibility Study of 10 July 2025 puts it in these words [R16]. "There is a well-established downstream refining process and terms in place for PGM concentrate within South Africa. Most smelters processing the concentrate from the Eastern and Western Limbs are situated in Rustenburg, with almost all the concentrator product in the area being transported by truck to Rustenburg. The Project PGM concentrates are expected to be processed at one of these facilities. Initial talks have been undertaken with smelters owners." The phrase "terms in place" refers to standard commercial terms existing in that market, not to terms agreed with Southern Palladium. The company has never described the talks themselves as anything more than initial. Every later mention of offtake is forward looking. The metallurgical announcement says the 1.2 per cent mass pull "is expected to enhance the Company's offtake pathway to large-scale smelters and refineries on the Western Limb" [R15]. The June 2026 quarterly says the result "opens the pathway towards a smelter and refining offtake arrangements" [R4]. The 2025 Annual Report lists progressing offtake arrangements as a priority for 2026 [R2]. Bridge Street, whose analyst holds shares in the company, puts it this way [R7]. "The single most valuable disclosure between now and the DFS would be a signed offtake agreement with a Western Limb smelter." Payability, and where the assumed number comes from A mine sells concentrate to a smelter, and the smelter pays for a proportion of the metal in it rather than all of it. That proportion is called payability, and it is agreed in the contract. The Pre-Feasibility Study's financial model assumes 85 per cent for platinum and palladium, 84 per cent for rhodium and gold, 56 per cent for ruthenium, 54 per cent for iridium and 100 per cent for the chrome concentrate [R8]. Evolution Capital's model uses the same figures, metal for metal [R18]. Those figures come from other companies' contracts. The study says the payabilities "were benchmarked from various other mines selling PGM concentrates through a third-party refiner", and the JORC table says they were "benchmarked based on numerous offtake agreements for PGM concentrates on the Bushveld Complex" [R8]. So the assumption rests on terms other companies have been given. Nobody has offered Southern Palladium terms. Bridge Street use 80 per cent, and explain why [R17]. "We have used conservative 80% payabilities for Pt and Pd in our valuation reflecting discussions held with third party concentrate suppliers." Five percentage points of payability on platinum and palladium, the two largest metals by weight, moves every economic number that depends on it. The gap between 80 and 85 per cent is there because no contract has been signed, and it closes when one is. An 86 per cent payability also appears in the Pre-Feasibility Study, and it is worth separating from the 85 per cent. The 86 per cent is the flat rate used in the resource pay-limit screen, which tests whether the rock has reasonable prospects for eventual economic extraction [R8]. The 85 per cent is the figure used in the financial model. The two do different jobs. Chrome is a penalty in the concentrate and a product on its own Chromite is refractory, which means it is hard to melt. In a conventional platinum smelting furnace, chrome in the feed raises the temperature at which the charge melts and can build up a high-melting layer on the furnace floor, which limits how much the furnace can treat. Smelters therefore set limits on how much Cr2O3 they will accept and price accordingly. Bengwenyama's test concentrates came in at 3.2 and 3.9 per cent Cr2O3 against the company's own target of under 2.5 per cent [R15]. Bridge Street, in a report otherwise strongly positive, calls this "one of the very few negatives we can find in this announcement", noting that "the Cr2O3 content of the concentrate is higher than the spec. required by smelters (3.2% vs 2.5%)" [R19]. The company says the same thing itself, and FIG. 06 is where it says it. The grades along the top are the good news. The list on the right is headed issues to close, and the first line of it is optimising the Cr2O3 content to meet commercial concentrate specifications. FIG. 06 Concentrate grades from the definitive feasibility study test work, and the company's own list of what is still open. Reproduced from the Southern Palladium investor presentation of 2 September 2026, slide 17 [R36]. Samso could not obtain a primary technical source for an industry-wide chrome limit, so no general figure is claimed here. The company's own target of under 2.5 per cent is the verified number, and the test concentrates came in 0.7 and 1.4 percentage points above it. That is the negative case in full, and it applies to one of two products. The same July 2026 test work that left 3.2 per cent chrome in the PGM concentrate recovered 85.6 per cent of the chromite into a separate concentrate grading 42.2 per cent Cr2O3, which is the grade the South African market trades [R15]. Section 3.00 sets out what that did to the shape of the mine and to the price assumption behind it. Commercially the chrome concentrate is the simpler of the two products. The Pre-Feasibility Study assumes 100 per cent payability on it, against 85 per cent on platinum and palladium, so the study expects none of the chrome revenue to be lost to a buyer's payment terms [R8]. The mining right expressly covers chrome, one of seven minerals granted [R3]. At full Stage 2 production the company expects about a million tonnes of it a year, and it delayed the definitive feasibility study by a quarter specifically to capture that value in the plant design [R3]. FIG. 07 is the market that concentrate would go into. On the sources the company cites, South Africa holds about 72 per cent of the world's chromite resources and produced 23 of the 44 million tonnes shown for 2025. A chrome producer in South Africa is selling into the deepest chrome market there is. FIG. 07 Where the world's chromite is, and where it is mined. Reproduced from the Southern Palladium investor presentation of 2 September 2026, slide 16 [R36]. The slide attributes the resource split to the International Chromium Development Association and the production split to the United States Geological Survey. It does not state the unit on the production chart. Samso has read it as tonnes and has not checked either series against its primary source. There is also a reason to think the two chromes move together rather than against each other. The spirals between flotation stages and the second chromite flotation stage that make the chrome concentrate are the same equipment that takes chrome out of the PGM concentrate [R15]. On that reading, recovering more chrome as product and getting Cr2O3 below 2.5 per cent are one job rather than two competing ones, which would explain why the company says the grades can be lowered in the commercial application. That is Samso's inference from the flowsheet description. The company does not put it that way and it has not been demonstrated at scale. At a 1.2 per cent mass pull and 443.6 grams per tonne on a 4E basis, Bengwenyama would truck a small tonnage of very high grade concentrate. That is an attractive parcel for a smelter with limited furnace capacity, because the metal per tonne of furnace feed is high. The chrome content counts against Bengwenyama in that negotiation and the grade counts for it. Neither product has an agreement behind it yet, and that is a step the company has not reached rather than one it has failed. Offtake terms are normally struck once a buyer can see a settled concentrate specification and a credible delivery date, and the July 2026 work is the first time Bengwenyama has had the first of those. What can be said today is that the chrome concentrate already meets a traded specification and the PGM concentrate does not yet, and that closing that gap is a plant design question the company has said it intends to answer. How the neighbouring mines get their concentrate smelted Every mine near Bengwenyama has a smelting route. TABLE 03 sets out what Samso could verify about each of them. Two Rivers is the one case Samso verified in a primary document. It has a life-of-mine agreement with Impala Refining Services, and Implats owns 46 per cent of Two Rivers [R20]. The concentrate has a contracted buyer because the smelter owner is a shareholder in the mine. Modikwa is owned equally by Valterra Platinum and African Rainbow Minerals, and Valterra owns smelters. Mototolo is wholly owned by Valterra. Booysendal is Northam's, and Northam owns the Zondereinde smelter. Marula is Implats'. Southern Palladium has no smelter shareholder. It would be a new entrant asking for capacity on commercial terms, which is a harder negotiation than the one its neighbours had. Waterberg, the cautionary comparison Platinum Group Metals Ltd's Waterberg project holds 18.7 million ounces and has been looking for a route to market for years. Bridge Street's February 2025 note describes Waterberg's difficulty [R17]. "Ivanplats (IVN TSX) has succeeded in securing capacity (with Northam) for Stage 1 concentrate volumes and part of Stage 2. This appears to have locked the Waterberg project of Platinum Group Metals (PTM TSX) out of the South African market for the time being." That is a broker's characterisation rather than a company's or a regulator's. The verifiable record is as follows. Northam concluded a ten-year offtake agreement with Ivanplats on 9 December 2021 for about 20,000 tonnes a year of Platreef concentrate [R21]. Platinum Group Metals' discussions with Implats broke down in 2022, after which the company investigated building its own smelter [R22]. In January 2025 it was still "in discussion with several South African smelter operators" while also pursuing a memorandum of understanding for a smelter in Saudi Arabia, which would need South African export permits for unrefined concentrate [R23]. Bridge Street, who are bullish on Southern Palladium, draw a distinction here that Samso thinks is fair. Waterberg is a low grade, base metal heavy Northern Limb concentrate. Bengwenyama would be a high grade, low volume UG2 concentrate. A smelter would look at the two differently. So Waterberg is evidence that a large resource does not guarantee a buyer, and it is not evidence that Bengwenyama will meet the same difficulty. Is South African smelter capacity actually full? The claim that South African smelter capacity is full is asserted often. Samso could not verify it from any smelter owner's own statement. No published statement from Valterra Platinum, Implats, Sibanye-Stillwater or Northam that Samso found says its capacity is full or fully contracted. Valterra Platinum's own guidance gives the one hard number available. In its half-year results presentation of 29 July 2026, Valterra guides purchases of third-party concentrate of 0.9 to 1.1 million ounces on a 6E basis for 2026, against 1.1 million ounces in 2025 [R24]. So a large third-party purchase market exists, and the guided volume is flat to lower year on year. Analyst commentary supports the claim without proving it. Miningmx reported in October 2024 that a shortage of refining capacity constrains platinum group metal production, and that Platreef's first phase concentrate could not have been processed by Northam at that time despite the signed offtake, because capacity would not have been available had material been supplied then rather than the following year [R25]. The route to a smelter, and the conditions a buyer would set What Bengwenyama would produce is a high grade UG2 concentrate, and that is relevant to who might buy it. Smelting and refining in South Africa sits with four companies. Impala Platinum, Sibanye Stillwater, Northam Platinum and Valterra Platinum run the furnaces and the refineries, and most of that capacity is on the Western Limb of the Bushveld Complex around Rustenburg. Bengwenyama is on the Eastern Limb, so concentrate would travel across the complex. The Eastern Limb mines already make that journey. A smelter with limited furnace capacity wants metal per tonne of feed, and this section has already made that point from the flowsheet. The same point works from the buyer's side. A high grade UG2 concentrate is the sort of parcel an integrated producer with a full furnace would look at, because it lifts the metal in the feed without adding much tonnage. What would a buyer want to see before opening a detailed negotiation? On Samso's reading of how these deals are normally done, two things. A granted mining right, so the seller can actually deliver. And metallurgical work at definitive feasibility study level, so the concentrate specification is settled enough to price. Both now exist. The metallurgical results were released on 9 July 2026 and the mining right was granted on 7 August 2026, and section 6.00 sets out what the grant does and does not cover. That is Samso's reasoning and not a company statement. Southern Palladium has not announced a condition set by any buyer, has not named a counterparty and has not put a date on a negotiation. What can be said is that the two things a buyer would most obviously want to see are now in place, where twelve months ago neither was. The verdict below does not move, because an argument about sequence is not an agreement. THE COUNTERARGUMENT Southern Palladium would be selling a small tonnage of high grade concentrate into a market with four integrated buyers, none of which owns any part of it, at a chrome content above the level it has told the market it is targeting, with no agreement and no quoted price. The sequencing argument above is Samso's and not the company's. Nobody has said that a granted mining right and a definitive feasibility study level metallurgical result unlock a negotiation. It is a reasonable reading of how these deals are done. It is still a reading, and the company has been describing offtake talks in the same forward looking language for two years. The commercial terms are the part nobody has seen. Payability, treatment charges, refining charges and the penalty for chrome above specification are the four numbers that decide what a tonne of Bengwenyama concentrate is worth. None of them has been quoted to this company. The financial model uses 85 per cent payability taken from other companies' contracts. Bridge Street use 80 and explain why. A new entrant with no smelter shareholding, asking four companies that each have their own ore to feed, is not the party holding the leverage in that negotiation. Chrome cuts both ways. The chrome that makes the second product is the same material a smelter will not want in the first one. The company says lowering one and recovering the other are the same job. That has not been shown at scale. Every other test in this piece can be answered from a document. This one cannot be answered until an offtake agreement is signed. Verdict on test five. Unresolved. No offtake agreement exists and no price has been quoted. Nothing has gone wrong here, the company has simply not secured a buyer yet, and this is the test that decides the project. 5.00 — TEST SIX, A RESERVE RATHER THAN A RESOURCE What the declared reserve covers A Mineral Resource says the metal is in the ground. An Ore Reserve says somebody has tested whether that metal can be dug up at a profit, applying the mining, processing, economic, legal and environmental factors involved. Southern Palladium declared a maiden Probable Ore Reserve on 23 October 2024 of 31.72 million tonnes at 6.17 grams per tonne on a 6E basis, for 6.29 million ounces [R8]. The reserve covers the UG2 reef only. The company's own notes state that no Inferred Resources were included in it, and that it was estimated using a 6E basket price of US$1,557 an ounce before payabilities. Three qualifications have to be read against that number. There is no Proved Reserve. Everything is Probable, including the tonnes converted from Measured Resource. Probable is the lower of the two confidence categories. The 40 million ounce figure is a resource, not a reserve. The total Mineral Resource across UG2 and Merensky reefs is 40.25 million ounces [R26]. Roughly three quarters of that is Inferred, the lowest confidence category. A large part of the Inferred is in the Merensky Reef, which is not in the mine plan at all, because the Pre-Feasibility Study and the Optimised Pre-Feasibility Study are UG2 only. The company also notes that several historic drill holes in one area were not assayed for the minor platinum group elements, so part of the resource is stated on a 4E basis and part on 7E [R4]. The 40.25 million ounce total is therefore a mixed basis number, and the reserve of 6.29 million ounces is the part that has been through an economic test. FIG. 08 shows why so much of that resource is within reach of a mine plan. Both reefs run shallow and close to parallel across the section, the Merensky Reef comes to surface at the modelled subcrop toward the right, and the horst block in the middle is the shallowest part of the whole line. A horst block is a slab of ground that faulting has pushed up relative to the ground either side of it. FIG. 08 Section A to A prime, with the Merensky and UG2 reefs, the horst block and the modelled Merensky subcrop. Reproduced from the JORC Table 1 appendix to Southern Palladium's definitive feasibility study metallurgical announcement of 9 July 2026 [R15]. Reef thicknesses are exaggerated on the original, as its own legend states. The mine plan is bigger than the reserve. The Optimised Pre-Feasibility Study states that ore production over the first five years is 83 per cent from Measured and Indicated resources, rising to 92 per cent over the first ten years, and that Measured and Indicated account for 70 per cent of total life-of-mine ore production [R16]. The other 30 per cent is Inferred, which by definition cannot be in a reserve. The Stage 1 and Stage 2 plan mines 41.9 million tonnes against a declared reserve of 31.72 million tonnes. That is not unusual for a study at this stage, and the company discloses all of it. The 30 per cent of the mine life that is Inferred depends on drilling that has not yet been done. The reserve, and the metallurgy that has changed since The Ore Reserve has an effective date of 23 October 2024 and it has not been restated. The Optimised Pre-Feasibility Study of July 2025 restructured the mine plan into two stages without changing the reserve, and the June 2026 quarterly still carries the October 2024 figures unchanged [R4]. In the 22 months since, the metallurgical assumptions behind the reserve have changed. Recovery moved from 85.3 to 87.6 per cent, chromite recovery from 30 to 85.6 per cent, and the primary grind from 60 per cent passing 75 microns to 30 per cent. Reserves are calculated using recovery and cost assumptions, so a reserve struck on the old assumptions does not reflect the plant the company now proposes. That cuts one way worth naming. A reserve is an economic calculation, so a reserve struck on 30 per cent chromite recovery and a US$225 chrome price is a reserve calculated on the worse of the two sets of assumptions the company has published. On the July 2026 test work the same rock carries more revenue per tonne, which lowers the grade at which mining a block pays for itself. The 6.29 million ounces is therefore more likely to be a floor than a ceiling. That is a direction rather than a forecast, and the restatement arrives with the study. The company's position, stated in the mining right announcement, is that "all the material assumptions underpinning this NPV continue to apply and have not materially changed" [R3]. The same announcement says the definitive feasibility study is being delayed a quarter specifically to capture the extra chrome value in the plant design. Both statements appear in the same document, and they are hard to reconcile. The restated numbers arrive with the study in the first quarter of 2027. The January 2026 retraction On 22 January 2026 Southern Palladium lodged a "Retraction of Updated Forecast Financial Information" [R31]. It withdrew slide 9 of a presentation the company had given at the Future Minerals Forum a week earlier, which restated the project's economics using a basket price of about US$2,000 for every 6E ounce. The Optimised Pre-Feasibility Study uses US$1,557. The company also withdrew three peer comparison graphs from the same presentation. The reason is the ASX Listing Rules. A company may not publish updated forecast financial information for a project unless it comes from a study that has itself been released to the market. The economics had been restated on a higher metal price without a new study behind them. The company retracted the material within a week rather than defending it, which is the correct response. And the episode is why the Optimised Pre-Feasibility Study of July 2025 remains the only set of economics for this project that can be relied on, whatever the chromite recovery has since done to the numbers. Anyone quoting a Bengwenyama net present value at today's metal prices is restating economics without a study behind them, which is what the ASX stopped the company from doing. Who signed the reserve The Competent Person for the Mineral Resources is Uwe Engelmann, a director of Minxcon (Pty) Ltd. The Competent Person for the Ore Reserves is Daan van Heerden, also a director of Minxcon [R8]. Both are reported under the JORC Code, 2012 edition. Minxcon is the consultant that provides Southern Palladium's geological and technical work, and it authored the Pre-Feasibility Study, the Optimised Pre-Feasibility Study addendum and the resource and reserve estimates. Southern Palladium paid Minxcon about A$655,000 in the June 2026 quarter alone [R4]. PLATE 03 is the work all of that rests on. Thirty-three thousand metres of drilling, core logged by hand, and two Competent Persons who put their names to what it showed. PLATE 03 Drilling, core logging and the core yard at Bengwenyama. Photographs from the Southern Palladium investor presentation of 2 September 2026 [R36]. Samso cannot date them. Both Competent Persons hold a beneficial interest in Southern Palladium through shareholdings in Nicolas Daniel Resources Proprietary Limited, which is one of the company's largest shareholders at 11.83 per cent [R2]. Daan van Heerden is also a non-executive director of Southern Palladium. The company discloses these interests in every announcement, which is what the disclosure rules are for. Arrangements like this are common in small resource companies and nothing here is an allegation. Samso sets it out because an investor reading a reserve statement should know who signed it and what they own. Verdict on test six. Passed. A declared reserve exists and it is entirely Probable. The Stage 1 and Stage 2 mine plan mines 10.2 million tonnes more than the reserve, and the reserve will be restated on new metallurgy in 2027. COFFEE WITH SAMSO The same companies, on the record Coffee with Samso is the interview series. Management answering questions in long form, so the people behind a project can be judged alongside the numbers in a piece like this one. Free to watch. Companies commission an interview, and it stays free to everyone, permanently. Watch the interviews → 6.00 — TEST SEVEN, PERMITS The mining right, granted 7 August 2026 The Department of Mineral and Petroleum Resources granted the mining right on 7 August 2026 under section 23(1) of the Mineral and Petroleum Resources Development Act, and Southern Palladium announced it on 10 August [R3]. The right runs for an initial 30 years with a right to apply for renewal. It covers 5,280 hectares across Nooitverwacht 324 KT and Eerstegeluk 327 KT, and it permits the mining of platinum group metals, gold, copper, chrome, cobalt, silver and nickel. The application was submitted on 29 September 2023 and accepted on 17 October 2023 [R4]. From submission to grant took nearly three years. That delay is a large part of why a shallow, high grade, well located UG2 property was still undeveloped in 2026. In South African law, granted, executed and registered are three separate stages [R27]. The announcement uses the word "granted" throughout and says nothing about execution or registration. A right is accepted when the Regional Manager confirms the application complies. It is granted when the Minister approves it, at which point the applicant becomes a mining right holder but mining cannot start. It is executed when it is signed by the Regional Manager and the holder in front of a notary, which fixes the effective date and starts a one-year clock to commence mining. It is registered at the Mineral and Petroleum Titles Registration Office within 60 days of the effective date, and registration is what creates a limited real right capable of being sold or encumbered. Registration is the stage that decides whether the right can be used as security. Until it is registered, the right cannot be pledged, and a project financier lending against a mine wants security over the right to mine it. Southern Palladium has not announced execution or registration. Samso could not verify either. The conditions attached to the grant have not been published. Kalkine reported on 10 August 2026 that the company indicated it would disclose the full detail of the terms and conditions once its own assessment was complete [R28]. A South African mining right will ordinarily carry conditions covering the approved social and labour plan, the environmental management programme and commencement obligations. None of those conditions is in the public record, so Samso does not describe them. What is still outstanding TABLE 04 lists which permits are granted and which are not. Two licences have to be granted before any work can start on the ground. The board has approved starting box cut and decline development before the end of 2026, and the announcement says that approval is "subject to completion of the relevant waste management and water use permitting processes" [R3]. Neither the water use licence nor the waste management licence has been granted. The 2025 Annual Report said water use licence applications were scheduled to be completed and submitted by December 2025, which means they had not been lodged when that report was signed [R2]. FIG. 09 is the company's own schedule for the work those two licences gate. Site clearance, access roads and fencing fall in 2026, the box cut and portal support run through 2027, and decline development and plant construction carry past the right-hand edge of the chart. The chart stops at 2030 and does not mark a first production date. FIG. 09 The development pathway as the company schedules it, from site clearance to plant construction. Reproduced from the Southern Palladium investor presentation of 2 September 2026, slide 18 [R36]. A schedule is a plan, and this one predates the grant of the two outstanding licences. Bridge Street's view is that granting these "will be a formality" [R19]. That is an opinion from an interested party. What can be verified is that neither licence is granted, and that the company has made its own start date conditional on both. One further question cannot be closed. Under section 96 of the Act, a person whose rights or legitimate expectations have been materially and adversely affected may lodge an internal appeal against an administrative decision, within 30 days of becoming aware of it [R29]. Samso could not establish whether any appeal has been lodged against this grant, or when the 30 days started, because the period runs from awareness rather than from the date of the decision. This ground has a documented history of third parties objecting, so an investor is entitled to ask the company whether an appeal exists rather than assume none does. Verdict on test seven. Mostly passed. The mining right is granted. The water use and waste management licences are not, and the right has not been shown to be executed or registered, so a lender cannot yet take security over it. 7.00 — TEST EIGHT, BEING ABLE TO RAISE THE MONEY The funding requirement, and what the company has Part 1 said the eighth test works differently from the other seven. It describes market conditions on the day rather than any property of the project, and those conditions change without anything at Bengwenyama changing. TABLE 05 sets out the requirement against what the company has. The Appendix 5B for the June 2026 quarter reports zero loan facilities, zero credit standby arrangements, zero total financing facilities and zero undrawn facilities [R4]. The company has no debt and no committed source of debt. Cash at 30 June 2026 was A$23.879 million, and the quarterly's own estimate of funding available was 11.9 quarters. That figure is calculated against a study phase burn of about A$2 million a quarter. That figure measures how long the company can keep studying. It says nothing about how long it could keep building. Peak funding of US$279 million comes from the Optimised Pre-Feasibility Study of July 2025 [R16]. It is a project level number on a 100 per cent basis. The study builds it as capital of US$226 million, plus ramp-up operating cost and working capital of US$123 million, less US$69 million of revenue during ramp-up. Those components add to US$280 million against a published figure of US$279 million, so they are rounded. The capital estimate carries an accuracy range of plus or minus 15 to 25 per cent. FIG. 10 is how the company frames that requirement now, and it adds two things this piece has not used. The US$279 million is Stage 1 peak funding, which the slide puts at 38 per cent below the original Pre-Feasibility Study, and Stage 2 is shown as funded out of Stage 1 cash flow rather than out of a second raise. Stage 1 is put at a 21.8 per cent internal rate of return on its own and Stage 2 at 26.4 per cent, with all-in sustaining costs of US$969 an ounce on a 6E basis in Stage 1 and US$821 in Stage 2. FIG. 10 Staged development, with Stage 1 peak funding, all-in sustaining costs and the two stage returns. Reproduced from the Southern Palladium investor presentation of 2 September 2026, slide 11 [R36]. The slide refers back to the Optimised Pre-Feasibility Study of 10 July 2025 [R16] and states that the material assumptions behind that NPV have not changed, so it restates the July 2025 study rather than adding a new one. The US$857 million NPV shown uses the Pre-Feasibility Study price deck. Two qualifications apply to that number and they do not both point the same way. The estimate predates the July 2026 metallurgical rewrite, so the plant it costs is not the plant now proposed, and a plant with a second product in it is not automatically a cheaper one. Against that, the same rewrite added a revenue stream the funding calculation does not count, so the money to be raised is being measured against a mine the company no longer intends to build. Separately, it is a 100 per cent figure while Southern Palladium owns 70 per cent of MUM, and nothing Samso read says how the community's 30 per cent share of the funding is to be met. What the dilution looks like Bridge Street models the funding explicitly, which is more than most research does. Their numbers give the size of the raise. Their August 2026 note assumes A$200 million of project debt and A$230 million of new equity raised at A$1.40 a share, against roughly 133 million shares on issue [R7]. A$230 million at A$1.40 is about 164 million new shares. That is more new shares than the 133 million currently on issue. The assumed raise price has fallen with the share price, from A$2.00 in February 2026 to A$1.67 in April and A$1.40 in July [R30]. Bridge Street say so themselves, that the July change "has contributed significant dilution compared with our last valuation where we assumed equity would be raised at $1.67/share" [R19]. That method is circular. Because the assumed raise price tracks the market price, a lower share price produces more assumed shares and a lower value per share. A per-share valuation built this way falls automatically when the stock falls, without anything happening at the project, so it is not a minimum value. Southern Palladium's own share price has ranged from A$0.195 in May 2025 to A$2.71 in February 2026, and Bridge Street's note has it at A$1.76 in August 2026 [R7]. The orebody did not change over that period. One thing has changed in how the company presents this, and it is worth separating from what has changed in the project. Nothing has been raised and no facility has been signed. What the 2 September presentation does is split the requirement in two, put Stage 1 at US$279 million and describe Stage 2 as funded out of Stage 1 cash flow rather than a second raise [R36]. If that holds, the dilution modelled above is one event rather than two. It also rests on Stage 1 reaching steady state and generating that cash, which on the company's own schedule in FIG. 09 is a question for 2029 or 2030. Verdict on test eight. Open. The company needs to raise roughly 1.7 times its market capitalisation. It has no debt facility, no offtake and a mining right not yet in registrable form. Part 1 said this test would be decided by market conditions rather than by the project, and that is still the case. SUPPORT SAMSO Independent, organic, and free to read We produce independent organic content and need your support to keep that going. Every contribution covers the cost of research, keeps the site running, and buys the time to write the next piece. A$10, A$20, A$30 or an amount you choose One time, weekly, monthly or yearly A donation buys nothing and unlocks nothing. It keeps the work going. Support Samso → 8.00 — WHAT THE RESEARCH SAYS, AND WHO PAID FOR IT Who covers this company, and who pays for the coverage Samso holds reports from two research houses on Southern Palladium, Bridge Street Capital Partners and Evolution Capital. Both disclose that they were paid by the company for the work. Samso has not found unpaid coverage of this company, and has not attempted to establish that none exists. Bridge Street's disclosure is the most extensive, and they set it out in every report [R7]. "He currently owns shares in SPD... BSCP are Corporate Advisors to SPD and may receive fees from SPD for services provided." They were also lead manager to the initial public offering and to the capital raisings since. Evolution Capital disclose that they "engaged and received a financial benefit by the company", which may have included corporate advisory, capital raising or underwriting work [R18]. The disclosures do not make the research wrong. Bridge Street have followed this project since before it listed, have been to the site, and publish their assumptions in more detail than most houses do. Samso quotes the disclosures rather than describing them, because the exact wording carries the information. The two headline valuation numbers, and why they differ Evolution Capital published a 12-month price target of A$4.80 in January 2026. Bridge Street's August 2026 note carries a valuation above A$9 a share. Those two figures look like a disagreement. They are not the same kind of number. Evolution builds its figure in three layers and sets each one out [R18]. It starts with an unrisked net asset value attributable to Southern Palladium of A$1,685 million. It applies an 85 per cent probability of development to get A$1,433 million. It adds net cash and subtracts the present value of corporate costs to reach A$1,452 million, then divides by 168 million shares on a pro-forma basis, which assumes two staged equity raisings. That gives the risked net asset value of A$8.65 a share. The A$4.80 target is 0.55 times that figure, a market multiple Evolution describes as covering liquidity, timing, funding and residual delivery risk. Those are two different discounts, not the same one applied twice. The 85 per cent asks whether the mine gets built. The 0.55 asks what the market will pay for a developer before it is built. Between them the two houses put a figure on the same asset anywhere from A$4.80 to above A$9. That gap is not a disagreement about the rock. It is a disagreement about how much to take off for everything that is not the rock, and it is the only place in this piece where the discount becomes visible. Evolution takes off 15 per cent for the chance the mine is never built, then takes off a further 45 per cent for what it calls liquidity, timing, funding and residual delivery risk [R18]. Bridge Street takes off neither and assumes the project is fully funded [R7]. Neither house uses the word jurisdiction anywhere in its reasoning. Samso looked for a published figure attributing any part of a South African discount to the country itself and could not find one. FIG. 11 is the company's own answer to that spread, and it is the plainest thing Southern Palladium has published about the discount. It puts itself at A$6.4 for every attributable resource ounce on a 4E basis, against Chalice Mining at A$26.9 and Bravo Mining at A$19.9, and its own grade is the highest on the chart. Read it for what it is. The peer set is the company's choice, the vertical axis is logarithmic so the gaps look smaller than they are, and the note on the slide says the calculations use data sourced from Evolution Capital, which is one of the two houses this section is about. FIG. 11 Enterprise value per attributable resource ounce against resource grade, on the company's own peer set. Reproduced from the Southern Palladium investor presentation of 2 September 2026, slide 8 [R36]. The slide states that the calculations are based on data sourced from Evolution Capital and company reports. The vertical axis is logarithmic. SAMSO TAKE What a reader should do with a discount nobody has itemised. Evolution takes off 45 per cent for liquidity, timing, funding and delivery, on top of the 15 per cent it takes off for the chance the mine is never built. Bridge Street applies nothing at all. Neither house names jurisdiction, and yet jurisdiction is the first thing most investors say about a South African project. Samso's view is that a discount applied for reasons nobody has written down is precisely where an investor should do their own work rather than accept the number. That may mean the discount is fair. It may mean it is where the opportunity is. This is Samso's opinion and not a finding, and it is not a recommendation to buy, hold or sell anything. Section 10.00 has the fuller argument in Noel's own words. Bridge Street publishes no rating and no price target at all. Its explanation is that "we are always reluctant to use our valuations as share price targets. There are too many variables at play" [R7]. Its figure carries no probability of development and assumes the project is fully funded. So the two net asset values are not a like-for-like pair either, and this is where a reader is most likely to be misled. Evolution's A$8.65 is already risked at 85 per cent. Its unrisked equivalent is about A$10.10 a share, which is Samso's arithmetic on Evolution's own components. Bridge Street's February 2026 figure is A$8.00 [R30]. So the gap between the two houses is about a quarter, not the 8 per cent the two headline numbers appear to show. The share counts differ as well. Evolution assumes 168 million shares after two staged raisings. Bridge Street counts about 297 million against the 133 million on issue today. Section 7.00 sets out what that means. What moved the numbers Bridge Street publishes its own valuation history with the price deck behind each step, which is more than most houses do and makes the next part checkable [R19]. The project valuation went from US$943 million in August 2025 to US$1,933 million in July 2026. The first move was price. In February 2026 the house lifted platinum from US$1,300 to US$1,700 an ounce, palladium from US$1,200 to US$1,400 and rhodium from US$7,500 to US$8,000, and assumed a weaker rand. That took the valuation to about US$1,360 million on the price deck alone [R30]. Everything after that was the project. The first round of definitive feasibility metallurgical results took it to US$1,810 million in April 2026, and the second round took it to US$1,933 million in July. So roughly 40 per cent of the total increase is metal prices and roughly 60 per cent is the metallurgy described in section 3.00. One claim does not survive the house's own table. The February 2026 note says three times that its price assumptions are at a significant discount to spot [R30]. In July 2026 Bridge Street published a run at spot prices alongside its house deck, and the spot run gives US$1,688 million against US$1,933 million [R19]. So by July the house deck was worth about 14 per cent more than spot rather than less. Two things drive that. The deck carried platinum and palladium above spot in July, and it assumes 17.5 rand to the United States dollar against a spot rate of 16.33. Bengwenyama's costs are almost all in rand, so a weaker assumed rand lowers the dollar cost base and lifts the valuation. In fairness to the house, the July and August notes drop the discount-to-spot language and simply state the assumptions against the spot price of the day. The claim that does not hold is the February one. Neither house publishes an annual cash flow schedule, so neither valuation can be added up and checked from outside. That is normal for research on a developer, and it does mean any single valuation figure is one modeller's view. 9.00 — THE VERDICT The scorecard across both parts TABLE 06 sets out the scorecard across both parts. Six of the eight have an answer supported by a public document. Grade, the orebody, capital intensity, the flowsheet, the reserve and the permits are all answered. On permits, the mining right is granted. The water use and waste management licences are still outstanding, and no work can start on the ground until both are issued. Two tests are not answered, and it is worth being precise about what that means. Test five has no offtake and test eight has no funding package. Neither is a finding against the project. Both are steps a developer takes late, and both are the kind of step that is either done or not done rather than done well or badly. There is no offtake agreement, no memorandum of understanding, no term sheet and no quoted payability, and the payability behind the project economics is benchmarked from what other companies on the Bushveld have been paid. That is a reasonable place to start and it is still an assumption rather than a price. What the scorecard does not show is the part of the project that has no line on it. The chrome is in none of the published economics. The reserve was struck on the lower of the two recovery assumptions the company has released. And the discount between the two research valuations is wider than the gap in their views of the orebody. Those three are where the value in Bengwenyama has yet to be priced, and none of them can be settled until the definitive feasibility study lands. THE CASE FOR Bengwenyama is one of the last shallow, outcropping, high grade sections of the UG2 on the Eastern Limb, in a district with sealed roads, grid power, water and a workforce that has mined this reef for a century. The mining right is granted for 30 years over seven minerals. A Probable ore reserve of 6.29 million ounces has been declared. The flowsheet is conventional and the company says so expressly. The July 2026 test work took chromite recovery from 30 per cent to 85.6 per cent and turned chrome from a by-product credit into a second revenue stream of about a million tonnes a year at full production. Capital intensity is about US$1,100 of capital per annual ounce, against about US$3,900 for the nearest ASX-listed comparison. The concentrate would be a small tonnage at more than 440 grams per tonne on a 4E basis. A smelter short of furnace capacity wants exactly that kind of parcel, because the metal per tonne of furnace feed is high. The community owns 30 per cent of the project and the land it is on, which removes the access and social licence problem that has stopped other South African projects. The ten years of litigation that kept this ground undeveloped ended in 2010 with the community winning, which is why the ground was available to buy. And the reason Bengwenyama is hard to compare with anything is the same reason it exists. The United States Geological Survey puts the largest platinum group metal resources and reserves in the Bushveld Complex, and South Africa holds roughly four fifths of the world's reserves [R32]. Investors outside South Africa have very few projects of this kind to measure this one against, and the nearest ASX comparison needs about three and a half times as much capital for each annual ounce. Power is cheaper too, with South African business electricity averaging about US$0.105 a kilowatt hour against about US$0.242 in Australia [R35], and no load shedding for more than a year to August 2026 [R33]. The four disclosures that would change the answer Four disclosures would materially change the answers above. These are the ones to watch for. A signed offtake agreement with a named smelter, with the payability and the chrome penalty disclosed. That answers test five and turns the payability assumption into a quoted number. Notarial execution and registration of the mining right at the titles office. That turns the grant into a real right a lender can take security over, and it starts the one-year period the holder has to commence mining. The water use licence and the waste management licence. The board has already approved starting the box cut and decline before the end of 2026, and neither can begin until both licences are granted. The definitive feasibility study, now due in the first quarter of 2027, carrying restated capital, restated operating costs, a restated reserve and the chrome revenue properly in the model rather than as a by-product credit. Southern Palladium's own share price has moved from A$0.195 to A$2.71 and back to A$1.76 across roughly fifteen months. None of those four items had been delivered at any point across those fifteen months. SAMSO TAKE Part 1 said the first seven tests describe the project and the eighth describes market conditions. Bengwenyama has now answered six of them from the public record. The two still open both need somebody outside the company to act. A smelter has to offer terms. Investors or a lender have to put up about 1.7 times the current market capitalisation. The company cannot make either happen, neither has a date on it, and neither has been refused. So Samso would put two questions to a holder rather than one. If the definitive feasibility study lands in the first quarter of 2027 with better numbers again, and there is still no offtake and no funding package, what has actually changed? And if it lands with chrome properly in the model for the first time, against a valuation discount that no research house has itemised, how would you know whether the price already reflects that or not? Samso does not answer either question here. DYOR. 10.00 — FROM WHERE I STAND Jurisdiction, the 30 per cent, and the offtake Every mining project has more than one problem to solve at the same time. Infrastructure, money, market forces, and the jurisdiction it happens to be in. Southern Palladium has all four. Jurisdiction is where most investors stop reading. The mainstream calls a place Tier 1 or it does not, and South Africa does not make that list. But there is plenty of evidence that projects work in jurisdictions the market calls difficult. What changes is not whether a project can be built. What changes is how much the people running it matter. In a hard jurisdiction, management and the skill set around them are everything. That brings me to the 70 and the 30. I do not read the 30 per cent as a hole in Southern Palladium's ownership. I read it as somebody in-country with a reason to want this project to work. A partner who lives on the ground, holds the land, and carries the same downside you do is not a complication. It is the thing that gets a project through the years when nothing is going right. Seventy per cent of a big elephant is a lot better than 100 per cent of an ant. South Africa is two ends of the spectrum in one country. It holds some of the richest PGE ground anywhere and some of the most economic deposits. It also carries the volatility and the constraints that scare most mining entrepreneurs off before they start. Both are true at once, and you do not get the first without accepting the second. Then there is the offtake. It is a concern. But these things take time, and it is a chicken and egg for everyone involved. The producer wants the offtake. The smelter wants certainty that the supply will actually be there, and it wants it at the most economical price it can get. Both sides are waiting on the other. Is it good to have an offtake before a project takes off? Absolutely. Does it have to exist in some form eventually? Yes. Does it always get signed before everything else does? No. So the question is not really whether the agreement is missing today. It is whether you believe this will evolve, and whether you believe they can raise the money. Management will have some play here that is not public. And the valuation looks drastically different the day an offtake is secured in some form. Raising the capital is the other half of the same decision. Good rock on its own does not build a mine. If you want more certainty before you decide, ask the company where the potential of an offtake sits. That is a fair question, and the answer tells you something either way. Then do your own research from that point. It will give you a level of certainty. And remember what this industry is. It does not take much for market dynamics to change overnight. What this Insight does is show the detail. The depth, the differences, and the variation in valuation. There is a lot in a project like this one to be studied, and that is the point of the piece. I am not telling you what to think about Southern Palladium. I am telling you there is enough here that you should not decide in five minutes. SAMSO ACCESS Some of the work does not appear here Samso holds back a set of Insights for members. Same method as this piece, the same sourcing and the same counterweights, on companies and sectors that do not make the public feed. A$20 a month · cancel any time See what is inside → 11.00 — REFERENCES AND SOURCES Each source below carries an identifier, [R1] to [R40]. The same identifier appears in the body text at the point that source is used, so any claim can be traced back to the document it came from. Every reproduced figure in this piece is the company's own artwork, and the credit under it names the announcement or presentation it came from, with the slide or page number. All tables are compiled by Samso from the sources named in each table caption. Where a calculation is Samso's own, the caption says so and the underlying figures are the companies' or the courts' own published numbers. [R1] Southern Palladium Limited, "Pre-quotation disclosure", ASX and JSE announcement dated 3 June 2022, released 6 June 2022. Source of the corporate chain, the acquisition terms, the clawback terms and the two project milestones. [R2] Southern Palladium Limited, "2025 Annual Report to Shareholders", financial year ended 30 June 2025. Source of the joint venture accounting note, the shareholder loan terms, the community economic interest, the social and labour plan statement, the substantial holders and the top twenty. [R3] Southern Palladium Limited, "Mining Right Granted for Southern Palladium's Bengwenyama World-class PGM-Chrome Project", ASX and JSE announcement, 10 August 2026. Source of the grant date, the term, the area, the minerals covered, the ownership statement and the early works conditions. [R4] Southern Palladium Limited, "Quarterly Activities Report and Appendix 5B for the quarter ended 30 June 2026", 30 July 2026. Source of the tenement schedule, the cash position, the financing facilities, the Jameson Cells assessment, and the reserve carried forward unchanged. [R5] Southern Palladium Limited, "Abridged Prospectus and Pre-Listing Statement, Secondary Listing on the JSE Main Board", 22 April 2022. Source of the Nurinox ownership and the community description. [R6] Bengwenyama Minerals (Pty) Ltd and Others v Genorah Resources (Pty) Ltd and Others (CCT 39/10) [2010] ZACC 26, Constitutional Court of South Africa, 30 November 2010. [R7] Bridge Street Capital Partners, "Southern Palladium Ltd, Mining Right granted. Conceptual valuation >A$9/share", August 2026. Bridge Street are corporate advisers to Southern Palladium, were lead manager to its initial public offering and to subsequent capital raisings, and the author holds shares in the company. [R8] Southern Palladium Limited, "Prefeasibility Study Results, Project NPV8 of USD1.059bn, Maiden JORC Ore Reserve of 6.29 million oz @6.17g/t PGM(6E)", ASX and JSE announcement, 28 October 2024, effective date 23 October 2024. [R9] Mineral and Petroleum Resources Development Act 28 of 2002, Republic of South Africa, sections 2, 11, 23, 96 and 100. [R10] Broad-Based Socio-Economic Empowerment Charter for the Mining and Minerals Industry, 2018, gazetted 27 September 2018, clause 2.1. Cited here through the summary published by Werksmans Attorneys. The gazette itself was not obtained. [R11] Minerals Council of South Africa v Minister of Mineral Resources and Energy and Thirteen Others (20341/19) [2021] ZAGPPHC 623, Gauteng Division of the High Court, Pretoria, 21 September 2021. [R12] Fasken, "Update, Mining Charter 2018 Judgment, DMRE Will Not Appeal", 26 November 2021. [R13] Werksmans Attorneys, "Section 11 of the Mineral and Petroleum Resources Development Act", quoting section 11(1) verbatim. [R14] White & Case, "Proposed overhaul of South African mining law regime", June 2025, on the draft Mineral Resources Development Bill published 20 May 2025 and corrected 9 June 2025. [R15] Southern Palladium Limited, "DFS metallurgical studies deliver outstanding PGM and Chromite recoveries", ASX and JSE announcement, 9 July 2026. Source of every recovery, grade, mass pull and flowsheet figure in section 3.00, and of the chrome content of the concentrate. [R16] Southern Palladium Limited, "Optimised Prefeasibility Study, Project NPV of US$857m with Improved Fundability Through Staged Development", ASX and JSE announcement, 10 July 2025. Source of the capital, peak funding, mine plan and resource classification percentages. [R17] Bridge Street Capital Partners, "Site visit confirms the quality of the Bengwenyama PGM project. Our NPV8 now US$1.1Bn", February 2025. Source of the 80 per cent payability assumption and the Waterberg characterisation. [R18] Evolution Capital, "Proven Bushveld Blueprint, De-Risked PGM Delivery, Southern Palladium Ltd", initiation report, 27 January 2026. Evolution Capital disclose that they received a financial benefit from the company. [R19] Bridge Street Capital Partners, "Significant value-add from final DFS metallurgical test work", July 2026. [R20] Impala Platinum, Two Rivers Platinum fact sheet, and Two Rivers Platinum company information, on the life-of-mine offtake with Impala Refining Services. [R21] Northam Platinum Holdings Limited, ten-year concentrate offtake agreement with Ivanplats, announced 9 December 2021. [R22] Miningmx, "PTM to go it alone on Waterberg palladium smelter after offtake talks with Implats stall", 23 March 2022. [R23] Polity, "Waterberg discussing concentrate offtake with South African smelter operators amid Saudi mulling", 15 January 2025. [R24] Valterra Platinum, "H1 2026 Results Presentation", 29 July 2026. Source of the purchase of concentrate guidance. [R25] Miningmx, "Structurally inflexible, the PGM sector can only watch and wait", 21 October 2024. [R26] Southern Palladium Limited, "Bengwenyama Mineral Resource Update, Total (UG2 & MR) Mineral Resource now 40.25Moz", JSE SENS announcement, 23 October 2024. [R27] Werksmans Attorneys, "Mining terminology, the difference between accepted, granted, executed and registered", 26 April 2023. [R28] Kalkine, "Southern Palladium (ASX:SPD) surges as Bengwenyama Mining Right lands", 10 August 2026. [R29] Cliffe Dekker Hofmeyr, "How to appeal an administrative decision by the DMRE, a discussion of the amended regulations", 3 February 2021, on section 96 of the Act and regulation 74. [R30] Bridge Street Capital Partners, "Our project value now +US$1.3bn delivering an A$8/share valuation for SPD", February 2026, and the published valuation and price deck history table reproduced in the July 2026 report. [R31] Southern Palladium Limited, "Retraction of Updated Forecast Financial Information", ASX and JSE announcement, 22 January 2026. Source of the retraction of the restated basket price economics and the three peer comparison graphs, and of the third quarter 2026 definitive feasibility study timing then current. [R32] United States Geological Survey, "Mineral Commodity Summaries 2026, Platinum-Group Metals", February 2026. Source of the statement that the largest resources and reserves are in the Bushveld Complex in South Africa, of South Africa's 63,000,000 kilograms of reserves against a world total of more than 76,000,000, of the description of Russia as the world's leading producer of mined palladium, and of the 2024 mine production figures for palladium and platinum. No Russian reserves figure is taken from this table. [R33] Jonathan Hoyle, "Good news for the 174-year-old industry in South Africa employing 472,000 people", BusinessTech, 23 August 2026. Source of the statement that Eskom has not implemented load shedding for over a year, of the energy availability factor of about 82 per cent on 24 July 2026, of industrial electricity prices having roughly doubled over the past decade, and of the shift by large users to self generation. Secondary source, reported rather than verified against Eskom's own operating data. [R34] Eskom Holdings SOC Ltd, "2026/2027 Tariff Increase", Eskom Distribution. Source of the approved standard tariff increase of 8.76 per cent for Eskom direct customers effective 1 April 2026, and 9.01 per cent for local authority tariffs effective 1 July 2026. [R35] GlobalPetrolPrices.com, electricity prices for businesses, 2023 to 2026 average, world average stated for the second quarter of 2026. Source of South Africa at about US$0.105 a kilowatt hour, Australia at about US$0.242, and a world average of about US$0.164. A commercial price aggregator rather than a regulator, and its business category is not the Megaflex or equivalent tariff a mine would negotiate. Reported, not verified. See F-21. [R36] Southern Palladium Limited, "Investor Presentation", ASX and JSE announcement, 2 September 2026. Source of the locality map, the processing flowsheet, the definitive feasibility study against Pre-Feasibility Study metallurgy comparison, the chrome co-product sizing, the global chromite supply charts, the concentrate grades and the issues to close list, the development schedule, the staged development panel carrying Stage 1 peak funding at 38 per cent below the Pre-Feasibility Study, and the enterprise value per attributable ounce chart. Also the source of the site, community and drilling photographs. [R37] Eskom Holdings SOC Ltd, "Winter outlook for power grid again predicts no loadshedding, system stability enables key energy security decisions to be updated", media statement, 22 April 2026. Source of the surplus of about 6 gigawatts of peak capacity forecast for the 2026 winter and of the 14 gigawatt unplanned outage stress case. [R38] Transnet SOC Ltd, network access agreements concluded with eleven private train operating companies, announced 13 May 2026 and reported by SAnews and Engineering News. Source of the eleven operators and of about 24 million tonnes of additional freight capacity. [R39] S&P Global Ratings sovereign rating actions on South Africa, November 2025 and 29 May 2026, as recorded in National Treasury of the Republic of South Africa media statements. Source of the BB foreign currency and BB+ local currency ratings, the positive outlook, the affirmation of both in May 2026, and the first upgrade in about twenty one years. [R40] National Treasury of the Republic of South Africa, "South Africa exits the FATF greylist on 24 October 2025", media statement, 24 October 2025. Source of the removal date and of the description of the grey list as the Financial Action Task Force register of countries under increased monitoring. Cover photo: Chromite (GeoDIL number - 1403), Darla Sondrol. CC0, via Wikimedia Commons (https://commons.wikimedia.org/wiki/File:Chromite_(GeoDIL_number_-_1403).jpg). Reproduced photograph, shown faded under a navy overlay on the article thumbnail. The Samso Way – Seek the Research Here at Samso, we pride ourselves on delivering content for investors that is independent and informed by over three decades of experience in the industry. Our content is well-researched and is only created if I see merit in discussing the company's story. Our mission is simple: cut through the noise and spotlight what matters—genuine stories, grounded insights, and real opportunity. Investors can explore our four core platforms: Coffee with Samso Samso Insights Samso News Samso Research There may be numerous paths to success in investing, but the common thread among successful individuals is that they remain committed to making informed decisions. Equip yourself with the right knowledge and tools, and you will be well on your way to achieving your financial goals. Most importantly, investors need to be absolutely diligent in understanding their own risk-reward tolerance and capabilities. Never bite off more than you can chew. As they say, Rome wasn't built in a day, and the Great Wall stood because it took centuries to complete. The Samso Philosophy: Stay curious. Stay sharp. And remember—digging deeper always uncovers the real value. In Life, there is no such thing as a Free Lunch. Never bite off more than you can chew is my parting comment. Happy Investing, and the only four-letter word you need to know is DYOR. To support our independent work, please head over to our Support Page and give us a helping hand in any of the ways listed. This is a new initiative for the Samso Platform, and it was always the concept of Samso when we started this journey in 2018. Disclaimer This article is general information only. It is not financial product advice and it does not take into account any reader's objectives, financial situation or needs. It contains no price targets, no valuations and no recommendation to buy, hold or sell any security. Companies named in this article are used as historical illustrations of events that have already occurred and no view is expressed on their current merits. Readers should consider obtaining advice from a licensed financial adviser and should conduct their own research before making any investment decision. Samso and its associates may hold interests in companies mentioned. Read full Disclaimer. About Samso Samso is a trusted platform that equips dedicated investors with up-to-date industry knowledge and insights from top CEOs and thought leaders. By staying informed on business advancements and market trends, investors can enhance their financial decisions through a combination of expert guidance and their own research. Coffee with Samso, Samso Insights, Samso News and Samso Research are four ways into the same work. Interviews, long-form research, short takes on announcements and company-specific studies. Every one of them is about how to think about investing rather than what to buy. Samso Insights | www.samso.com.au | An Investor Lens on ASX-Listed Companies

A Harvard risk manager, a Nobel laureate and a Bell Labs mathematician have all reached the same conclusion about deciding how much to own. Samso went looking for what the professionals actually say, and set it against the realities of the Australian resources market. Samso Insights Investor Education Investor's Playbook Samso Investing Series