Brightstar Resources Limited (ASX: BTR), Part 2 of 3

Contents

5 The Mineral Resource and the Inferred Question	9 Building the Laverton Mill
6 Reading the Ore Reserve	10 The Balance Sheet, the Bond and the Revenue Gap
7 Sandstone	16 References and Sources
8 The 2026 Drilling, and How to Read the Headline Intercepts	17 Disclaimer

5 The Mineral Resource and the Inferred Question

A Mineral Resource is a geological estimate of how much metal is in the ground, made by a Competent Person under the JORC Code. A Competent Person is a named, qualified geologist or engineer who signs the estimate off and is personally accountable for it. It is reported in three confidence categories, Measured, Indicated and Inferred, in that order of confidence. An Ore Reserve, which Section 6 defines in full, is the part of that resource a company has shown it can mine and sell at a profit.

There are two current group resource numbers and both are live. The 15 July 2026 announcement gives 4.46 million ounces. The FY2026 Annual Report, released seven weeks later on 4 September 2026, states the group resource as at 30 June 2026 at 82.0 million tonnes at 1.5 grams of gold per tonne for 4.03 million contained ounces. The annual report figure predates the July Sandstone upgrade, so both are correct as at their stated dates. The company's presentations use the July number and describe the group as "over 4.5Moz". We use 4.46 million ounces and name its date.

What the confidence split means

Measured material is 52,000 ounces, which is 1.2 per cent of the group resource, and all of it sits at Laverton, spread across four deposits, being Beta at 19,000 ounces, Lord Byron at 17,000, Second Fortune 12,000 and Fish 4,000.[R2] Two of those four, Second Fortune and Fish, are the company's underground mines. Second Fortune is scheduled to cease during the September 2026 quarter and go to care and maintenance. Fish was paused in March 2026 and is held in operational readiness, with stoping to recommence in early calendar 2027.[R3] Lord Byron is not a small deposit, at 267,000 ounces in total, but only 17,000 of them are in the highest confidence class. Inferred material is 2.53 million ounces, which is 56.8 per cent.

An Inferred Mineral Resource is defined in the JORC Code as an estimate where quantity and grade are estimated on limited evidence and sampling. It cannot be converted directly into an Ore Reserve and it cannot be used in a feasibility study as if it were reserve. It is a real geological estimate made by a Competent Person, and a large share of Inferred material in an early-stage district is entirely normal. Sandstone has been drilled by Brightstar for barely two years.

The number does tell an investor where the work still is. Converting Inferred to Indicated takes drilling on a tighter spacing. Brightstar states its own classification rule for Sandstone plainly, that Indicated is defined "through areas predominantly defined by drilling at 20m to 40m spacing (or less)" and Inferred "where drill spacing varied from 40m up to 80m". That is a description rather than a rule. The JORC Code prescribes no drill spacing for any category. Classification is the Competent Person's judgement, and the company lists spacing as one criterion among geological continuity, confidence in the volume models, data quality, lode continuity and estimation parameters.[R2] Its own numbers show why the distinction matters.

Bull Oak is drilled on nominal 30 metre spacing, which sits inside the stated Indicated band, and it is still 86 per cent Inferred. The actual grid varies across the district by a factor of ten, from 10 metres to more than 100. At the tight end, Shillington and Vanguard Main sit on about 20 metres by 20 metres. At the loose end, Duplex is on 50 metre sections with holes 40 metres apart and Havilah is on 40 by 40, and both of those deposits are entirely Inferred. Lord Henry is drilled at 20 metres by 20 metres only to an average depth of 75 metres.

At Lord Henry, below about 75 metres, the spacing opens out to between 40 metres and more than 100 metres, so the deeper part of that deposit is more thinly drilled than the shallow part. The gap between the two categories at Sandstone is broadly a halving of drill spacing, and the company has been drilling hard to close it. The 15 July 2026 resource announcement reports "over 146,000m of drilling" at Sandstone since Brightstar acquired the ground, of which "approximately 60%" was incorporated into that estimate.

The cut-off grade, unconstrained reporting and bulk density

The Sandstone cut-off grade was lowered in the July 2026 update, from 0.5 grams to 0.4 grams per tonne for open pit material. The company discloses this and provides the like-for-like comparison, which is good practice. "As a like-for-like comparison, total Mineral Resources at the previous cut-off grade (0.5g/t Au) would now total: 62.8Mt @ 1.4g/t Au for 2.8Moz." So of the growth from 2.4 to 2.9 million ounces, about 70,000 ounces is the cut-off change and the rest is drilling.

Most of the Sandstone resource is reported unconstrained, meaning it is not confined inside an optimised pit shell. Again the company says so, and gives its reason. "Considering the shallow depths of the deposits, Brightstar has chosen to report most of the deposits unconstrained at a 0.4g/t Au cut-off." For shallow, flat-lying deposits that is a defensible reporting choice.

Unconstrained reporting does not mean the company skipped the test. Brightstar states that it ran conventional open pit optimisations across the Sandstone deposits, and that at a conservative gold price the resulting shells "encapsulate much of the interpreted mineralisation at each deposit (except for Bull Oak and TMH)".[R2] So the work was done. What unconstrained reporting means for a reader is narrower. The published tonnes were not clipped back to those shells, so the reported figure is not the same as the in-pit figure. The company has not published the constrained number alongside it, and that is the comparison we would like to see.

One more feature of the resource needs stating, and it is a normal disclosure and not a criticism. Bulk density, which is the factor that converts a volume of rock into tonnes, is measured in fresh rock and assumed in the oxide at several deposits. DFS 2.0 states that the values applied to oxide and transitional material at Aspacia, Lady Shenton and Yunndaga "are assumed", and that "[d]ensity values at CTW have been assumed and are based on values applied at neighbouring deposits". Lord Byron is the exception, with over 1,500 determinations. Oxide is the first material mined and the material the early cash flow depends on, so a reader should know its tonnage factor is largely inferred from elsewhere. Two of the three funded pits are affected.

Cork Tree Well, whose densities DFS 2.0 says have been "assumed and are based on values applied at neighbouring deposits", carries 104,000 ounces or 30 per cent of the reserve.[R1] Lady Shenton's oxide and transitional densities are also assumed, and it carries another 117,000. Only Lord Byron's are measured, from 1,567 determinations.

The two exceptions to unconstrained reporting are the two deposits now reported partly as underground, and they run the other way. Bull Oak is reported on an open pit basis only to 165 metres vertical and at a higher cut-off of 0.8 grams of gold per tonne below that, because below that depth the company considers it an underground opportunity. Two Mile Hill is reported to 140 metres vertical, and below that at a cut-off of 0.73 grams of gold per tonne confined within the tonalite, for the same reason. Together those two deposits hold 1.05 million ounces, 37 per cent of Sandstone, and the parts of them below about 150 metres are being treated as underground material, not open pit.

Depletion, and the Fish and Second Fortune lines

The resource statements carry the note that "Mineral Resources are depleted for historical mining". During FY2026 Brightstar mined 244,500 tonnes at 3.0 grams of gold per tonne for about 24,000 contained ounces from Second Fortune and Fish. The Fish and Second Fortune resource lines are identical in the 19 May 2025, 30 June 2026 and 15 July 2026 tables, at 376,000 tonnes for 49,000 ounces and 92,000 tonnes for 40,000 ounces respectively.

Second Fortune mined more tonnes in FY2026 than its resource line carries, at 128,400 tonnes against 92,000. That on its own is expected rather than odd. Second Fortune is a narrow vein reported at 13.4 grams of gold per tonne and mined at 3.08 grams of gold per tonne, and the difference is waste rock dug out with it. The ounces are the test. Second Fortune produced 12,700 contained ounces during the year and Fish 11,200, and both resource lines are unchanged across the statements of 19 May 2025, 30 June 2026 and 15 July 2026.[R3] The published figures are what they are, and both the annual report and the July announcement carry the depletion note. Both figures stand as released. The reconciliation is the company's to make. Combined, the two lines are 89,000 ounces of a 4.46 million ounce resource, so nothing in the investment case turns on it.

6 Reading the Ore Reserve

An Ore Reserve is the part of the Mineral Resource in Section 5 that a company has shown it can mine and sell at a profit, after applying mining costs, processing costs, recoveries, dilution and a gold price. It is the highest-confidence category in the JORC Code and it is the number a lender looks at.

Four deposits carry a reserve. Proved material is 15,000 ounces, or 4.3 per cent of the reserve ounces, and all of it sits at Lord Byron. The company's own figure, 4.4 per cent, is struck on tonnes. The company states this itself in its JORC Table 1. "[O]nly one of the deposits (Lord Byron) contains some Measured Ore Resources which converted to Proven Ore Reserves."

Before the arithmetic, the assumptions. A reserve is a resource with the mining costs and losses applied to it, and DFS 2.0 publishes the ones it used. All three open pits carry 10 per cent mining dilution, which is waste rock that gets dug and milled along with the ore and drags the grade down. Mining recovery, the share of the ore that actually reaches the mill rather than being left behind or lost, is 95 per cent at Lady Shenton and 90 per cent at Cork Tree Well and Lord Byron. The minimum mining width in each pit is 20 metres. Underground, the ore drives carry 14 per cent development dilution and the stope designs carry a 0.5 metre allowance for wall rock breaking into the void. Mining recovery there is 95 per cent in the stopes and 100 per cent in development.[R1]

The cut-off grades are published too, and they are marginal ones, set after the pit has already paid for the digging. As the study puts it, the decision on a given truckload treats the mining cost as already spent. At Lady Shenton they run 0.63 to 0.69 grams of gold per tonne, above the 0.5 grams the resource is reported at. At Cork Tree Well and Lord Byron they run 0.37 to 0.51 grams of gold per tonne, below it.[R1] Both are defensible on the study's own logic. A reader comparing a resource tonne with a reserve tonne should know they are not struck on the same basis.

Conversion is reserve ounces divided by resource ounces, and it is a rough score for how much of a company's gold has been proven up. Which number you get depends entirely on which resource you divide into, and for Brightstar the two answers are very far apart.

The 351,000 ounce reserve sits on the Laverton and Menzies ground the feasibility study covers, and nowhere else. DFS 2.0 publishes the resource on exactly that ground, across seven deposits, in its own Table 4, at 26.17 million tonnes at 1.6 grams of gold per tonne for 1,351,000 ounces.[R1] Divide the reserve into that and conversion is 26.0 per cent. Divide it into the whole company, 4.46 million ounces, and it is 7.9 per cent.

The difference is Sandstone. Sandstone has never had a feasibility study, and without one it cannot carry an Ore Reserve at all. So its 2.87 million ounces sit in the bottom of the second sum with no possibility of appearing in the top of it. Both figures are true and they answer different questions. The 26 per cent says how well the company converted on the ground where it did the work. The 7.9 per cent says how much of the company is still unproven, which is also worth knowing. This workup gives both wherever the number appears, and Section 13 sets the two against the peer group.

On the studied ground, measured against Measured and Indicated ounces alone, which is the fairer comparison again because Inferred material cannot convert to reserve, conversion is 52.4 per cent. Against group Measured and Indicated it is 18.2 per cent.

How much of the mine plan is reserve, and how much is not

The company sets this out plainly, which is to its credit.

The company describes this split as "Measured (3%), Indicated (70%) and Inferred Mineral Resources (27%)". Those percentages are correct on a tonnage basis. On contained ounces, which is what actually turns into revenue, Inferred is 30.5 per cent rather than 27 per cent, because the Inferred material carries a higher grade than the Indicated material. Both figures are right on their own basis. Which basis is being quoted changes the number by three and a half points.

Brightstar publishes the required cautionary statement, and it publishes it in full. It is reproduced as released.

"Investors are cautioned that there is a low level of geological confidence in Inferred Mineral Resources and there is no certainty that further drilling will result in the determination of Measured or Indicated Mineral Resources, or that the production target will be realised. Of the Mineral Resources scheduled for development and extraction in this Production Target during the payback period, approximately 70% is classified as Measured or Indicated and 30% as Inferred over the initial Base Case 17-month payback period following mill commissioning. The financial viability of the Project is not dependent on the inclusion of Inferred Resources." Brightstar Resources, "Updated Goldfields Feasibility Study", ASX announcement, 29 January 2026

Three observations on that statement. The first is in the quote itself. The 17-month payback is about 30 per cent Inferred on the company's own account. So both the headline payback and the headline 74 per cent rate of return depend on material that cannot be put in an Ore Reserve, in exactly the window where the balance sheet is tightest.

The second is that the Inferred proportion is not spread evenly across the mine life. The study's own annual chart gives it year by year, and the shape matters more than the average does.

Figure 4. The plan leans hardest on its lowest-confidence material in its biggest year. Inferred material is 27 per cent of the ore mined across the whole plan, but 41 per cent of it in FY2030, which is the peak production year at 79,000 ounces. Both the annual percentages and the plan average are as published by the company on a tonnage basis. The company's chart covers FY2028 to FY2033, from the first full year of commercial production, while the 27 per cent plan average includes the part-year FY2027. Source: Samso, from Brightstar Resources, "Updated Goldfields Feasibility Study", ASX announcement, 29 January 2026[R1].

FY2030 is the peak production year in the plan at 79,000 ounces, and it is the year most dependent on the lowest-confidence material. A reader who takes the 27 per cent average as the exposure will understate what the middle of the plan is carrying.

The third is the company's own claim alongside the disclosure, that "the financial viability of the Project is not dependent on the inclusion of Inferred Resources". The claim is a strong one and it may well be right. The study does describe how it tested it. Its JORC Table 1 says the mine plan was built first with the Inferred material in, then re-optimised "with inferred material treated as waste and economics re-checked based on the revenues from indicated resources", with the viability of the reserve then re-checked.[R1] That is the right test. What is missing is the answer. No NPV, no rate of return and no cash flow from that re-optimised case is published anywhere, and the disclosure sits under the underground mining assumptions rather than under the project as a whole.

A reader who wants to size the margin cannot do it from what is published. Samso's view is that a reserve-only sensitivity would strengthen rather than weaken the investment case, because it would put a floor under it. It is the disclosure we would most like to see added.

The reserve was prepared in-house

The Competent Person for the Ore Reserve is Andrew Rich, and the annual report describes him accurately as "an Executive Director and shareholder of Brightstar Resources Limited". The underground mine design, scheduling and costing were "completed internally by Brightstar". The JORC Table 1 records, under audits and reviews of the Ore Reserve, "None have been completed yet."

Nothing there breaches the JORC Code. An in-house Competent Person is permitted and common, and the interest is disclosed exactly as it should be. Section 12 returns to the incentive question, because Mr Rich's short-term and long-term incentives both carry reserve growth targets.

7 Sandstone

Sandstone is what the case for holding this company beyond six years rests on, and it carries the least completed work of the three hubs.

What is established

The resource is 69.3 million tonnes at 1.3 grams of gold per tonne for 2.87 million contained ounces, across fifteen deposits, as at 15 July 2026. Indicated material is 1.14 million ounces. There is no Measured material at all. Brightstar's own Sandstone ground carried 1.5 million ounces before the Aurumin scheme, being 1.05 million ounces on the ground that came with Alto Metals and 0.5 million ounces at Montague from Gateway Mining[R5]. Aurumin took the combined total to 2.44 million ounces in November 2025[R6], and drilling has since taken it to 2.87 million ounces, which the company states as an 18 per cent increase. Most of the growth to date came from buying rather than drilling. The July 2026 update lifted the Indicated component by more than 100 per cent, and that part is drilling.

The deposits are spread across a district rather than concentrated, which is what a consolidation produces. Two Mile Hill at 731,000 ounces is the anchor and it is roughly a quarter of Sandstone on its own.

Figure 5. Sandstone is a district rather than a deposit. The nine camps labelled here hold 2,205,000 of the 2,868,000 ounces in the Sandstone resource, and Two Mile Hill, the largest, holds roughly a quarter of the total on its own. The remaining ounces sit in deposits this map does not label, which Table 12 lists in full. Resources are as at 15 July 2026. Source: Brightstar Resources, "High Grade Exploration Success Continues at Sandstone", ASX announcement, 7 September 2026, Figure 1.[R16]

What is not established

Sandstone has no Ore Reserve, no processing plant, no completed feasibility study of any level, and no final investment decision. The Pre-Feasibility Study and a maiden Ore Reserve are due in the December 2026 quarter, having been targeted first for mid-2026 and then for the September 2026 quarter. A Definitive Feasibility Study and a final investment decision are aspired to in late calendar 2027 or early calendar 2028, and construction in the second half of calendar 2027. Section 11 sets out how those two dates sit against each other and why they are aspirations rather than targets.

The metallurgical test work at Sandstone is thin. All of the test work disclosed in the July 2026 resource announcement is third-party legacy work. Middle Island Resources tested "a suite of samples, representing select Sandstone deposits" at ALS in 2020 and recovered between 85.1 and 98.3 per cent of the gold contained in the samples after 24 hours, "across grind size P80 of 125, 106, and 75um". That range is not a like-for-like number. It mixes a coarse grind with a fine one, and the company names neither the deposits tested nor which grind produced which result. Gateway Mining submitted 15 samples from Montague-Boulder in 2023, of which 12 were leach-tested. That is one deposit out of the 15 at Sandstone.

Recoveries came back above 90 per cent of the contained gold within 24 hours, with no grind size disclosed. Brightstar states that it "has engaged Independent Metallurgical Operations Pty Ltd (IMO) to conduct Definitive Feasibility Study-level metallurgical test-work on material collected from each deposit". The tense there is future and no results have been reported.

So for Two Mile Hill at 731,000 ounces and Bull Oak at 322,000 ounces, the two deposits the company now reports partly on an underground basis, there is no published metallurgical recovery. Both are deposits the company now describes as underground opportunities below 140 to 165 metres. Recovery is one of the two or three numbers that decide whether a plant works, and it has not been published for the two largest deposits.

The shallow ounces claim

The company has said repeatedly since January 2026 that 1.6 million ounces of the Sandstone resource sit within 150 metres of surface, and 1.2 million ounces within 100 metres.[R11] That claim carries weight because shallow ounces mean a low strip ratio and a cheap open pit.

We could not find a depth-sliced resource table supporting it anywhere in the 37 source documents. It appears as a bullet point and in managing director commentary, and never in a JORC table. In the same January 2026 announcement the claim also appears as "90% (1.6Moz) of the Mineral Resource located within the top 150m from surface". Against the 2.4 million ounce resource of the day, 1.6 divided by 2.4 is 67 per cent. The 90 per cent works only against a different denominator disclosed two paragraphs later, the "40Mt @ 1.4g/t Au for ~1.8Moz Au" being assessed for open pit mining. Both figures are as published in the same document.

The claim is also now stale. It was made against a 2.4 million ounce resource, the resource is 2.87 million ounces, and the growth since has been substantially at depth. Bull Oak grew from 90,000 to 322,000 ounces largely by drilling below the old shallow limit, and the Two Mile Hill extension work is at around 700 metres.

None of that means Sandstone is not shallow. There is real evidence that it is, and it comes from the company's own technical discussion rather than its headlines. The mill cut-off grades used in the pit optimisations range from 0.26 grams a tonne for oxide material to 0.46 grams a tonne for fresh rock. These are marginal cut-off grades, which the company defines in DFS 2.0 as costs "excluding mining cost, as the decision between the truck load being ore or waste considers the mining cost as a sunk cost". In plain terms, a 0.26 gram cut-off does not mean 0.26 gram material is worth mining. It means it is worth putting through the mill once it is already on the truck.

Even read that way it is a very low number, and it only works on soft, near-surface material that is cheap to mine and cheap to treat. Bull Oak is "intensely kaolinised to clay plus quartz to a depth of approximately 60m below surface", meaning the rock has weathered to soft clay and is cheap to dig. The lateritic soils next to it, being the iron-rich crust that forms over long weathering in this country, "contain significant gold mineralisation". Drill results back it up. Havilah returned 12 metres at 2.24 grams from 3 metres and 4 metres at 11.1 grams from 6 metres. Indomitable East returned 7 metres at 1.53 grams of gold per tonne from surface.

Samso's view is that the shallow open pit case at Sandstone is real and is supported by the technical disclosure. The specific "1.6Moz within 150m" figure is the weakest way to make that case, and an investor is better served by the cut-off grades and the near-surface intercepts, which are sourced and checkable.

Why Samso puts no value on Sandstone

Rovira was asked directly for a value on Sandstone in May 2026, and declined to give one. "Oh, it's a very delicate question there… I'll be careful with how I answer that… I'm not going to describe any value to it." He then moved to the resource's size and shallow depth, and from there to peer market capitalisations.

He was right to be careful, and Samso will not put a number on it either. The PFS has to deliver four things for the story to work. It needs a maiden Ore Reserve, a plant size, a mine life and a capital cost, and it needs metallurgy on the deposits that will feed the plant. That study is due within the next four months.

8 The 2026 Drilling, and How to Read the Headline Intercepts

Brightstar has run a heavy drilling programme at Sandstone through 2026, up to six rigs at times, and it has produced the longest intercepts in Brightstar's own record. Those results deserve a careful reading, because the headlines and the assay tables underneath them tell slightly different stories, and both are worth understanding.

The drill types, in plain English

Four drilling configurations appear throughout these announcements and they are not interchangeable. The hole prefixes in every table that follows are built from these codes.

Two Mile Hill and the tonalite

Two Mile Hill is the largest target drilled in 2026, and the reason is the rock it sits in. The deposit sits in a tonalite. A tonalite is a coarse-grained intrusive igneous rock made mostly of plagioclase feldspar and quartz, with little or no alkali feldspar. In plain terms, it is a pale granite-like rock that cooled slowly deep underground.

Tonalite is hard and brittle, so when the region was squeezed it fractured rather than smeared, and those fractures filled with gold-bearing quartz. The company puts the model well. "The Two Mile Hill tonalite is a felsic intrusive stock that has acted as a geological trap for gold mineralisation. Similar felsic intrusive bodies are seen throughout the Yilgarn and are known to host major gold deposits, including Thunderbox, Gruyere, Kanowna Belle, and King of the Hills."

That comparison is fair as far as it goes. Those are four large, real Western Australian gold mines and all of them are hosted in or around intrusive bodies. All four also sit in the Eastern Goldfields, while Sandstone sits in the Southern Cross domain on the other side of the craton. Naming them is a statement about the deposit model, not about Two Mile Hill's size, and the company does not claim otherwise.

Two mineralisation styles are recognised. Shillington-type sits in banded iron formation. Two Mile Hill-type sits inside the tonalite as sheeted, high-grade quartz veins dipping shallowly to the north-east, "set within broader envelopes of lower-grade mineralisation throughout the tonalite". That last phrase governs how the intercepts below should be read.

Figure 6. Two Mile Hill in plan view. The tonalite intrusion is shaded pink, banded iron formation blue and basalt green. The section line marked runs through TMHRCD26020, the deep hole discussed below. Figure 7 is a second section about 75 metres north, through TMHDD26004. Source: Brightstar Resources, "Spectacular 305m @ 1.8g/t Au in Sandstone extension drilling", ASX announcement, 3 August 2026, Figure 6.[R13]

Figure 7. Two Mile Hill in cross section, on the section through TMHDD26004. The tonalite widens with depth and the July 2026 resource block model stops well above the deepest drilling. The widened zone at the bottom of the section is drawn by the company with dashed boundaries and question marks on both flanks. It is an interpretation from two holes rather than a drilled outline. Yellow stars mark logged visible gold in that hole, which had no assays at the time of publication. Source: Brightstar Resources, "Significant visible gold in 500m mineralised intersection", ASX announcement, 25 August 2026, Figure 5.[R15]

What the drilling found

The drill spacing that carries the resource classification puts these intercepts in place. The 15 July 2026 resource[R2] says Two Mile Hill is defined by drilling on nominal 20 metre sections with holes 20 to 40 metres apart on each section, angled at 60 degrees to the west, "with average depth of 120m". Every intercept in the table above starts below 200 metres downhole and most start below 250. The company is explicit that the deepest of this work, the Shirvington Zone, "sits outside of the existing Mineral Resource Estimate".

Most of these holes were infill rather than step-outs. The company gives the first aim of the programme as "Infill areas optimised for underground mining at a sufficient drill spacing to facilitate an upgrade to the higher confidence Indicated category".[R13] Infill is drilling inside ground already estimated, to raise the confidence category rather than to add ounces. The exception is TMHRCD26020, the 305.4 metre hole, drilled under the programme's second stated aim, to "Test the eastern contact of the tonalite at depth to confirm the thickness of the unit".[R13] The deepest work, at what the company calls the Shirvington Zone, sits outside the current resource, and it says so.[R14] Either way none of it is near-term mill feed and none of it is in the funded plan. Sandstone has no Ore Reserve and no mine schedule until the pre-feasibility study lands, which the company now expects in the December 2026 quarter.

The tables above and below use gram-metres, which is the intercept length multiplied by its grade. It is the standard way to compare a long low-grade hit against a short high-grade one, and it is how we work out where the metal actually sits.

These are large intercepts and the grades are real. Three things need saying about how to read them, and none of them is an accusation.

These are downhole lengths. The company's own 25 August release shows how easily downhole length and true width slide into each other. Its headline reports "+500m thickness of mineralised Tonalite". Its managing director, on the same page, calls the same rock "downhole widths of 400-500m". He is right and the headline is not, and the difference between them is the whole of this section. The company says so in a line repeated in every 2026 exploration announcement. "True widths are not confirmed at this time although all drilling is planned perpendicular to interpreted strike of the target lodes at the time of drilling." Downhole length is simply how far the drill travelled through mineralised rock. True width is how thick the body actually is. If a hole cuts a lode at an angle rather than square on, the downhole length is longer than the true width, sometimes much longer.

Drilling perpendicular to strike fixes the direction the hole points on a map. It does not fix the angle at which the hole meets the dip of the rock, and that is what sets true width. At Two Mile Hill the two are different things. The high-grade veins dip shallowly to the north-east and most of the tonalite holes are inclined at about 55 to 61 degrees towards 270 degrees, which is close to square on the veins and a good angle to drill them. It is a poor angle on the tonalite itself. The company states that the tonalite dips about 78 degrees towards 281 degrees, so a hole plunging 58 degrees to the west runs at a shallow angle along the intrusion rather than across it.

The company makes the point for us without meaning to. Its 27 August announcement says the June programme infilled "a part of the ore body where the tonalite is approximately 80m thick", and the headline result from that programme is 225.7 metres downhole. The reported intercept length is roughly three times the thickness of the rock the company says hosts it.

TMHRCD26020 is the clearest case, because the company publishes both orientations. The hole was drilled at 81 degrees on an azimuth of 069 degrees, meaning it points 69 degrees east of north. The same JORC table has the tonalite dipping about 78 degrees towards 281. The hole and the intrusion therefore lean away from each other and the acute angle between them is around 20 degrees. That hole is running close to along the tonalite rather than across it. The company's reading of this hole is that it tripled the interpreted thickness of the intrusion, from about 60 metres to more than 180 metres. Its 25 August release widened that again, to more than 300 metres at 700 metres depth.[R13][R14] It is open that the interpretation is early, and its own cross section draws the widened zone with dashed boundaries and question marks on both flanks.

Working from the company's own geometry, 305.4 metres downhole at 81 degrees covers about 302 metres of vertical distance and only about 48 metres of horizontal run. It measures how far down the tonalite continues, which is useful to know. It does not measure how wide the tonalite is. This is not a quirk of one steep hole. The company describes the tonalite as dipping about 78 degrees towards 281 degrees, and the standard Two Mile Hill holes are drilled at 55 to 61 degrees to the west.[R13] Those meet the rock at about the same 19 or 20 degrees, so the stretch factor is roughly three across the whole table. The company's own numbers show it. A 225.7 metre intercept downhole through ground it describes as about 80 metres thick is a ratio of 2.8.

The cut-off rule behind the headline is worth setting out. Brightstar's JORC Table 1 says significant intercepts are reported above 1.0 gram per tonne with a maximum of two metres of internal dilution, "unless otherwise stated". The headline table in each announcement is captioned differently, above 0.5 grams of gold per tonne and unconstrained by internal dilution. Both are disclosed. A reader who reads only the headline and assumes the one-gram rule applies will have the wrong picture.

Most of the metal sits in a small part of the length. Every number below comes from the company's own published assay tables.

Figure 8. In four of Brightstar's 2026 headline intercepts, a single sample of one metre or less carries between a third and ninety per cent of the contained metal. Without that sample the remainder is still ore, at 1.1 to 1.4 grams of gold per tonne at Two Mile Hill, and it is a different rock from the one the headline describes. Source: Samso, calculated from the full assay tables in Brightstar Resources ASX announcements of 3 August[R13], 27 August[R14] and 7 September 2026[R16]. Gram-metres is length multiplied by grade, and all lengths are downhole lengths.

The announcement's own assay table breaks the 305.4 metres at 1.82 grams of gold per tonne down. About 95.6 metres of it reports above one gram per tonne, and those 95.6 metres carry roughly 90 per cent of the contained metal. The remaining 210 metres averages about 0.26 grams per tonne. That is below the 0.73 gram cut-off the company itself applies to Two Mile Hill tonalite at this depth, where the rock would have to be mined from underground rather than from a pit.[R2]

The 305.4 metres at 1.82 grams is a correctly length-weighted average and the company publishes the full table underneath it. The company's own body text is also more careful than its headline, describing the interval as "comprised of numerous higher-grade intervals within a wider lower grade tonalite envelope". The description is accurate. The interval is not 305 metres of uniform 1.8 gram a tonne rock. It is roughly 96 metres of ore-grade material inside a much wider envelope of sub-cut-off tonalite.

One further detail on TMHRCD26020, and the company reports it in full.

"The drillhole unexpectedly continued through tonalite to the end-of-hole depth of 657.3m, at which point the drill rig ran out of drill rods." Brightstar Resources, "Spectacular 305m @ 1.8g/t Au in Sandstone extension drilling", ASX announcement, 3 August 2026[R13]

The hole ended at 657.3 metres because the rig ran out of drill rods rather than because the tonalite ended. Nothing below that depth has been assayed.

Visible gold and the nugget effect

Brightstar drilled its first hole past a thousand metres in August 2026.[R15] TMHDD26003 reached 1,039.2 metres and TMHDD26004 reached 888.3 metres. Both ended in mineralised tonalite, and between them the geologists logged 53 separate occurrences of visible gold.

Visible gold in core draws attention, and it needs to be understood for what it is. It is a logging observation rather than a grade. The company publishes the standard caution and publishes it properly, that "visual estimates of mineral abundance should never be considered a proxy or substitute for laboratory analyses". It also gives the size, describing grains "typically <1mm in diameter but vary from pinhead to match-head size", and it estimates the percentage of gold in the mineralised zones at less than 0.01 per cent.

There is a second thing coarse visible gold tells a geologist, and it cuts the other way. When gold occurs as discrete visible grains rather than as fine particles spread evenly through the rock, sampling becomes harder. Two samples taken a metre apart in the same rock can assay very differently, depending on whether a grain happened to land in the sample. In geological jargon this is called the nugget effect. It does not mean the gold is not there. It means that any single very high assay may not repeat, and that a resource estimate has to apply top cuts to stop one sample dominating a block.

The evidence that this is live at Sandstone is in Brightstar's own data. Table 15 above shows single metres assaying 449 grams and 751 grams of gold per tonne carrying most of the metal in their intercepts. The company publishes photographs of the gold nuggets recovered from the 751 gram metre, which is good disclosure. It also states, correctly, that "no top cuts are applied to the assays when calculating intercepts". Leaving assays uncapped is normal and transparent for exploration reporting. It means every headline number above is uncapped, and the resource estimate built on the same data does cap them. The July 2026 statement says top cuts were applied to high grade outliers, domain by domain, chosen off log probability plots and histograms.[R2] What it does not publish is a single top-cut value.

For a deposit where the company photographs its gold nuggets, the top cuts applied at Two Mile Hill and Bull Oak could be added to the next statement.

Two things Brightstar does well here should be said, because the nugget effect is often reported as though it were only a negative. The company assays its RC samples by Photon Assay on a nominal 500 gram sub-sample. Photon Assay irradiates and counts gold across the whole sample without dissolving it, compared with a conventional fire assay that uses a 30 to 50 gram split. A 500 gram charge is ten to seventeen times the sample mass, and that is exactly the mitigation the JORC guidance points to for coarse gold. The mitigation has a limit, and the record does not settle where it falls. The same JORC table answers the question twice. Its sampling section says diamond core was "submitted to Intertek for analysis by fire assay".

Its assay quality section says "1m and 4m RC composite, and diamond drilling samples were assayed via the Photon Assay method at Intertek laboratory, Perth". Both sentences are in the same table of the same announcement. The Photon mitigation certainly covers the RC results, including the Bull Oak metre at 751 grams. Whether it covers the diamond tails at Two Mile Hill cannot be determined from what is published. The legacy Aurumin procedure also records that "where visible gold was observed a flush was passed through the core saw and a barren flush inserted in the sample sequence", which is the right practice for stopping gold carrying over between samples.

No screen fire assay programme has been reported at Sandstone. For a deposit with photographed gold nuggets in RC chips, screen fire assay is the industry-standard confirmation. The pulp, which is the finely ground powder the sample is crushed down to, is screened and the entire coarse fraction is fire-assayed. The only screen fire assaying described anywhere in the record is legacy work by Brightstar's predecessors at Laverton. The June 2025 study reports it, noting that "the presence of visible gold in the limited diamond core prompted the use of Screen Fire assay". The JORC Table 1 also does not respond to the Code's explicit prompt to explain coarse gold sampling, despite the same announcements publishing photographs of nuggets.

The 500 gram Photon charge substitutes for a good deal of this on the RC side, and Samso's view is that it is largely adequate there. Photon addresses the laboratory sub-sampling step and not the field splitting step. The company designs a 2 kilogram split off the rig-mounted cone splitter and takes the 500 gram Photon charge from that.[R13] The residual risk sits in that first split off the cyclone, and Photon does not reach it. Naming the gap is still the right thing to do, because it is the one check a reader would want before treating a 751 gram metre as representative.

What is open, and on what evidence

Open at depth means the drilling has not closed the mineralisation off below the deepest intercept. It is a useful thing to say, because it points at where the next work will go. It is also a phrase that implies mineralisation continues without meaning that anyone has evidence for it.

Applied to Brightstar, the claims vary in strength. They are worth separating.

Menzies is the best supported, and only at Yunndaga. Yunndaga is modelled to 334 metres vertical and the historical workings there reached 630 metres, so there is documented mineralisation nearly 300 metres below the modelled base. Lady Shenton is the other way round. Its model reaches 335 metres vertical while the historical underground workings stop at 200 metres and the old pit at 80 metres, so at Lady Shenton the case rests on the drilling rather than on the workings.

Bull Oak is well supported. Mineralisation has been intersected past 500 metres downhole and the company's own table describes the intrusion as having "a depth of at least 250m". Eight diamond holes are still pending assay.

Two Mile Hill below 700 metres is the strongest claim in the portfolio and it currently rests on the least evidence. Two holes ended in mineralised tonalite with visible gold logged to 907 metres, and the assays for both are due in late September 2026. Until those return, the Shirvington Zone is a logging observation.

Indomitable East is the weakest. Thirty-three RC holes for about 3,400 metres[R17] average roughly 103 metres a hole, and the deepest reported intercept runs from 128 to 139 metres. The company describes mineralisation as remaining "open, with significant growth opportunities at depth", and at that deposit depth has effectively not been tested, though it should be said that this programme was not meant to test it. The company describes it as 33 holes "designed to infill mineralisation within a portion of the current Mineral Resource, ensuring sufficient drill spacing for future MRE updates to support Indicated Mineral Resource classification". The depth claim is waiting on a programme that has not been run.

The same announcement also describes the deposit two ways, as a north-west trending sequence dipping steeply to the north in the body text and as "subvertical" with "an east-west strike" in its own JORC table. Both descriptions appear in the same release, and the orientation matters because it sets whether the holes cut the body squarely.

Not everything grew, and the company reports that too. The Two Mile Hill and Shillington resource fell from 755,000 ounces on 8 July 2026 to 731,000 ounces in the 15 July estimate after what the company called "rigorous geological interpretation and robust estimation parameters". It fell while the open pit cut-off was being lowered from 0.5 to 0.4 grams of gold per tonne, and a lower cut-off adds ounces on its own. So the reduction from the remodelling is larger than the 24,000 ounce difference on the face of it.[R2] Publishing it anyway is to the company's credit. Three holes at Lord Byron returned no significant intersection, two of them substantial step-outs to the east-southeast, and Brightstar published them along with a reinterpretation of the lode's dip and strike that explained why.[R12] Reporting failed holes and a corrected geological model is above-average practice and it should be credited.

9 Building the Laverton Mill

The contract

GR Engineering Services, ASX code GNG, is building the plant. It was named preferred contractor on 17 February 2026[R18] and the contract was executed on 26 May 2026 alongside the final investment decision.[R20] The contract sum is A$110 million on a lump-sum fixed-price basis. It covers engineering, procurement and construction of a 1.5 million tonne per annum CIL plant with a gravity circuit, engineered so it can later be expanded to 2.5 million tonnes.

DFS 2.0 did not factor the plant up from rules of thumb. It carried a lump-sum tender price. Its Table 17 prices the work under a column headed "Lump Sum Project Pricing", the section above it says the contractor has priced the defined scope "on a fixed and firm basis", and the flowsheet on the next page is credited to GR Engineering.[R1] Note that DFS 2.0 captions that table "Table 17: Processing Operating Costs Based on Material Type", repeating its Table 16 caption, while the table holds capital costs. Stated as published. The competitive front-end engineering process, in which rival contractors design and price the plant before one is chosen, closed in December 2025, a month before the study was published. Note also that DFS 2.0 applies a single plus or minus 15 per cent accuracy band to a capital estimate whose largest component is a firm lump sum price.[R1] The uncertainty in that A$133 million is not spread the way the label implies.

That uncertainty sits almost entirely in the A$15 million of non-processing infrastructure, in the owner's scope, and in whatever falls outside the contractor's defined scope. Note also that the A$11.8 million contingency Table 17 adds to the processing capital does not appear in Table 18's A$133 million sub-total, which is A$118 million of processing infrastructure plus A$15 million of non-processing infrastructure and nothing else. The study does not reconcile the two. Stated as published.[R1] The winning contractor was not named until 17 February 2026. So the A$110.9 million in the study and the A$110 million contract executed on 26 May 2026 are the same contractor's price for the same scope, quoted twice. The study still treats the figure as an estimate.

It applies the same plus or minus 15 per cent accuracy band to it as to everything else, and adds a 10 per cent contingency of A$11.8 million on top of the A$118.0 million processing total.[R1] The agreement between them is not independent corroboration, because it is one contractor's price quoted twice. What it does establish is that the plant line was priced by tender rather than factored up from rules of thumb. The price then survived four months and a binding contract without moving, which is a real and unusual strength in a first-time developer's study. It says nothing about the other A$361 million of life-of-mine capital, none of which was priced this way.

A lump-sum fixed-price engineering, procurement and construction contract, shortened to EPC, transfers the cost risk of the defined scope to the contractor. For a first-time builder it is the right structure, and it is the single biggest de-risking decision the company has made on the plant. It does not cover the larger number. Open pit mining is A$324 million of operating cost in DFS 2.0. The three open pits in the Ore Reserve, being Lord Byron, Cork Tree Well and Lady Shenton, carry A$242 million of capitalised mining between them, against A$110 million for the mill. None of it is under contract. The study costed it from "experienced consultants, mining contractors, and current rates from Brightstar operations" and said the fleet "will be supplied and operated by a reputable surface mining contractor who will be selected from a competitive tender process completed prior to mining".[R1]

As at 21 May 2026 the company still listed finalising the open pit mining contractor as an outstanding step, and no award has been announced since. Mobilisation for Lord Byron is targeted for the December 2026 quarter.[R3] So the de-risking achieved on the plant has not yet been achieved on the part of the project that costs five times as much. Rovira's own view of the contractor is short.

"They are genuinely the best in the space at building West Australian gold processing plants." Alex Rovira, Managing Director, Brightstar investor briefing webinar, July 2026[R34]

That is a view Samso cannot independently verify, and GR Engineering's current contract book, set out below, is consistent with it.

Two things about the contract are not disclosed. The company states, in managing director commentary and not in a contract summary, that the structure comes "with cost, schedule and performance guarantees". No liquidated damages regime, cap, performance bond, retention or parent company guarantee is described anywhere in the record. On a A$110 million lump-sum contract those are the terms that decide who carries an overrun.

The second is contractor loading, and this comes from GR Engineering's own disclosure rather than Brightstar's. In its FY2026 results, guidance and equity raising announcement of 24 August 2026[R19], GR Engineering reported FY2026 revenue of A$493.2 million and guided FY2027 revenue of A$825 million to A$850 million. That is an increase of 67 to 72 per cent. It also reported more than A$1.0 billion of new contract wins since 1 April 2026, and raised A$110 million of equity alongside the result. The same announcement lists its EPC contracts. Brightstar's A$110 million sits fifth by value, behind Develop Global at A$275 million, Ora Banda at A$233 million, BHP Iron Ore at A$230 million and Vault Minerals at A$155 million.

A contractor lifting revenue by roughly 70 per cent in one year and raising equity to fund the working capital is a contractor whose resourcing is stretched across the state. It is outside Brightstar's control and inside its critical path.

Figure 9. The Laverton plant site, published 9 September 2026, roughly three months into construction. The company publishes the image uncaptioned and unannotated, so the areas in it are not identified. The company publishes no percentage of completion, so the only schedules in the record are Figure 5 of the 2 July 2026 update and the same chart re-issued as Figure 8 of the 9 September 2026 update, drawn in quarters with no durations and no float. Source: Brightstar Resources, "Laverton Mill Construction Advancing on Schedule", ASX announcement, 9 September 2026.[R21]

The schedule, and whether it has moved

The first gold date has not moved. Every statement from February 2026 to 9 September 2026 says the June 2027 quarter. Nine statements over seven months with an unmoved date is a good record, and slipping schedules are the normal experience in this sector. Four of those nine have come since the final investment decision on 26 May 2026.

Two things temper it. The only schedules published are Figure 5 of the 2 July 2026 update and the same chart re-issued as Figure 8 of the 9 September 2026 update. Both are drawn in quarters, with no durations and no float, float being the spare time a task can absorb before it delays everything after it.[R22][R21] The processing plant bar runs up to the first gold marker in the June 2027 quarter with no separate commissioning or ramp-up band before it. At that resolution a slip of less than a quarter does not show. And no document publishes a weather or wet season allowance in the construction schedule. DFS 1.0 published Laverton's rainfall data, a 237 millimetre annual mean concentrated in the summer months.

DFS 2.0 dropped the climate section, and neither study carries it into a schedule float.[R4] Structural steel, mechanical and electrical installation are four of the nine items on the study's own critical path, and they run through the December 2026 to March 2027 storm season into a June 2027 first gold.[R1]

Two other dates did move, and both movements are modest. The final investment decision was targeted for late in the March 2026 quarter and was declared on 26 May 2026, roughly two months late, waiting on regulatory approvals and debt settlement. The company extended its early works agreement with GR Engineering to 31 May 2026 to hold the construction schedule while it waited, which is a sensible response. Mobilisation for the Lord Byron open pit has moved. The 21 May and 2 July 2026 documents both put site establishment in the September 2026 quarter. The 4 August and 9 September updates put mobilisation in the December 2026 quarter, with infrastructure work starting in October. That is a one-quarter slip, disclosed without being called one.

Where construction actually stands

The 9 September 2026 update reports bulk earthworks for the raw and process water ponds complete, and the CIL tank ring beams, slab and containment walls complete with the tank installation team mobilised to site. The primary crusher walls first lift and suspended floor are poured, the SAG mill plinths are poured and the thickener foundations are complete. The borefield bores are drilled and cased with the main header pipework delivered. One piece of mechanical equipment is on site, a vibrating screen. Twelve long-lead equipment packages had been ordered by 21 May 2026, including the SAG mill, variable speed drive, thickener, crushers and cyclones. The SAG mill and variable speed drive orders were placed under the early works agreement by 17 February 2026.[R18]

The language is consistently positive. "Advancing strongly", "progressing well", "tracking within the approved budget", "on schedule and within budget", and from Rovira, "disciplined execution through the civil construction and concrete phase which has ensured delivery to schedule".

No percentage of construction completion has been published, in this update or any previous one. The only percentage in the whole construction record is engineering design, reported at 60 per cent complete on 2 July 2026 and "on track to reach 90% completion in August". No document after 2 July confirms the 90 per cent milestone was reached.

That absence is worth naming without overstating it. Monthly percentage-complete reporting is not a listing requirement and plenty of builders do not provide it. For an investor it means there is currently no quantitative way to check the "on schedule" claim from outside the company. The observable facts are that the concrete is going in, the schedule has not moved in seven months, and one piece of mechanical equipment is on site nine and a half months before first gold. The plant needs its own power station, because there is no electricity grid at Laverton, and that scope item sits outside the lump sum and carries no published number.

DFS 2.0 lists a power station among the infrastructure it says is included in its capital and operating costs, and publishes no split for it.[R1] The company executed a long-term power purchase agreement with PWR Hybrid Solutions and an LNG supply agreement with EVOL LNG, both OCTA Group companies, in July 2026, and has secured the long lead equipment. It has published no value, term, tariff or gas price for either.[R25][R33] Power generation and diesel are two of the inputs the study names as driving its operating cost, and the study says the project is more sensitive to operating cost volatility than to capital cost volatility. So the cost of running the plant rests partly on a contract nobody outside the company has seen.

DFS 2.0 publishes the critical path, and four of its nine items are still ahead. The study lists them in order. Contract signing and the start of detailed design. Then securing the major equipment, being the long lead items. Then earthworks, civil and SMP contractor mobilisation to site. Then mill installation, construction of the tank farm, electrical mobilisation and electrical works. Commissioning is last.[R1] SMP is structural, mechanical and piping work, which is the erection of the steel and the tanks. Those are the phases where schedules usually slip.

The company's own two updates read differently on when that started. The 4 August update reports that in July "structural steel and platework for the primary crushing area commenced". The 9 September update lists "the first structural steel erection" in its look-ahead for the month.[R21] Both are as published.

What is approved and what is not

Two mining approval gaps sit in the record. No mining approval is disclosed anywhere for Cork Tree Well, which carries A$98 million of capitalised mining and is one of the three open pits delivering the bulk of Goldfields production. None is disclosed for Yunndaga either, which is scheduled to start in calendar 2027. Yunndaga is also a mine being planned inside ground that was mined out a century ago. The DFS plans mining to a depth of 235 metres below surface[R1], and the historical underground workings there reached 630 metres[R8]. The company addresses it, stating that "none of the known workings intersect the reserve mining shapes" as designed.

The word doing the work in that sentence is "known", because plans from 1896 to 1943 are rarely complete, and old voids, water and ground support are real schedule risks. That aside, the DFS JORC table notes that "Yunndaga exhibits some PAF material, which requires further follow up testwork", and adds that "the waste rock landform for Yunndaga has been designed appropriately to encapsulate any PAF material such that environmental risk is mitigated". PAF stands for potentially acid forming, meaning waste rock that can generate acid drainage if it is not managed, and it is a normal thing to encounter and a normal thing to design around. It is also a thing regulators look at closely.

No groundwater abstraction licence is mentioned anywhere in the source documents, while borefield drilling and pipework proceed. It may be granted, applied for or not required. It cannot be determined from what is published. The 15 September 2026 Mikado purchase adds, in the company's words, "additional footprint for water exploration and borefield infrastructure", which secures ground to look for water on[R35]. It is not a licence to take any.

The operating record so far

Brightstar has mined and sold gold through an ore purchase agreement with Genesis Minerals. It trucked ore from Second Fortune and Fish to Genesis Minerals' Laverton mill and sold it. Five campaigns ran from March 2025 to March 2026, totalling about 400,000 tonnes at 2.2 grams of gold per tonne for about 24,000 recovered ounces, and A$138 million of gross sales proceeds.[R23][R25]

That is an operating record, and it paid for things that are still on the ground. Rovira lists them. The development of the Fish mine, a camp of about 170 people at Fish, housing the mill construction crew now and intended for the Lord Byron and Fish mining workforces, ancillary infrastructure and haul roads. He calls it "a significant sunk cost and infrastructure advantage", and on the evidence that is fair.

One campaign went wrong and the company reported it. Parcel 4, processed in November 2025, ran 83,000 tonnes at 2.32 grams of gold per tonne and recovered 75 per cent of the gold that ore contained, against 91.3 per cent for the parcel before it. The explanation in the annual report is specific.[R3][R24] "The drop in recovery was related to an increase in the mineral Pyrrhotite, which consumed oxygen in the leach circuit which was subsequently addressed in the fifth parcel through milling practises and ore blending." The company quantified the shortfall at "underperformance of ~1,650 recovered ounces ($10.5 million lower revenue)", measured against its own forecast recoveries across mined production rather than against the previous parcel alone. Parcel 5 then recovered 89.1 per cent, so the fix worked.

Pyrrhotite is a reactive iron sulphide. In a cyanide leach it consumes the dissolved oxygen the gold dissolution reaction needs, and it also consumes free cyanide and generates sulphur species that can coat the gold surface, so recovery falls on two counts. It is a known problem with known remedies. The standard ones are pre-aeration, oxygen or peroxide addition and lead nitrate. None of those appears anywhere in Brightstar's record. The company reports that the problem was "subsequently addressed in the fifth parcel through milling practises and ore blending", which dilutes the problem instead of treating it. On the following 138,000 tonne parcel, through somebody else's mill, that worked, and the diagnosis and the disclosure were prompt. It is a point in favour of the company's operating capability.

The pyrrhotite question is worth carrying forward rather than closing off. The December 2025 quarterly names the cause as "the increased proportion of Pyrrhotite-bearing Fish ore in the blended parcel". Fish is a banded iron formation deposit that Brightstar's own resource documents describe as "sulphide facies interflow sediments" with quartz-magnetite-amphibole banding, so the pyrrhotite is part of the rock and not a one-off. Fish sits in the DFS 2.0 mine plan carrying an assumed recovery of 94 per cent of the gold contained in its fresh ore and A$18 million of capitalised mining. The one campaign in which Fish ore was a large share of the blend recovered 75 per cent. It is the only operating data point anybody has on this ore, and it sits nineteen points below the assumption. It is not a prediction about the Laverton plant.

Brightstar has never published an achieved all-in sustaining cost, an achieved C1 cost or any cost per ounce or per tonne for its own operations. Every AISC in every document is a DFS 2.0 forecast, being the group figure of A$2,998 an ounce and per-deposit all-in sustaining costs ranging from A$2,813 to A$3,516 an ounce. The company does publish quarterly cost tables in dollars alongside ounces sold, so the information is not hidden, but it never normalises them to a per-ounce figure.

Doing that arithmetic from the published quarterly tables gives cash operating costs plus royalties and sustaining capital of roughly A$4,000 to A$8,200 an ounce across the five quarters to March 2026. Those are Samso's calculations, not the company's, and they should be read with care. The mines being costed there are narrow-vein underground operations with a hundred-kilometre-plus unsealed haul to somebody else's mill. The DFS AISC of A$2,998 describes open pits feeding an owned mill on site. They are not the same business and the comparison is not apples to apples. The A$2,998 has no operating history behind it, and the first quarter of production will be the first time anyone can check it.

10 The Balance Sheet, the Bond and the Revenue Gap

What is in the bank

At 30 June 2026 Brightstar held A$122.0 million of cash and cash equivalents, made up of A$44.8 million at bank and A$77.2 million in term deposits. It held a further A$160.9 million of restricted cash, being bond proceeds held in escrow, which means a third party holds the money and releases it only when agreed conditions are met. The directors' report totals the two at A$283 million, against A$11.7 million a year earlier.

That is a well-funded balance sheet for a company of this size, and it is the result of the financing completed in February and March 2026.

The funding package

The bond terms are set out below.

It is interest-only for the first 18 months. It then amortises at 7.5 per cent of the issue amount per quarter for four quarters, then 10 per cent per quarter for five quarters, with a final 20 per cent of the principal repaid in one lump at maturity. At the company's own exchange rate of 0.70 US dollars to the Australian dollar that is about A$12.9 million a quarter, rising to about A$17.1 million. The 12.5 per cent coupon runs on top of it, against an operation planned to produce 75,000 ounces a year. The conversion is Samso's, at the company's rate. The first principal payment falls roughly two quarters after first gold. Security is senior over the Goldfields and Sandstone subsidiaries. There are three financial covenants.

The first is minimum liquidity of A$15 million and the second is an equity ratio of not less than 35 per cent, meaning shareholders' funds must be at least 35 per cent of total assets, at each half-year. The third is a leverage ratio of no more than 3.25 to 1 from 30 June 2028, meaning borrowings may not exceed 3.25 times annual operating earnings, falling to 2 to 1 from 31 March 2029. The company was in compliance at 30 June 2026. First drawdown from escrow was expected around October 2026, subject to a cost-to-complete test, which checks that the money still in hand covers the work still to be done at each drawdown.

There are no warrants, no royalties, no streaming and no mandatory hedging. The company states it plainly, that there are "no options, warrants, royalties or other dilutive price participation or production-linked instruments in the debt structure" and "no mandatory hedging instruments that cap gold price upside". Samso has not surveyed comparable development debt, so the view that this is better than the norm is offered as a view rather than a finding. Most first-time developers at this scale end up giving away a royalty, a stream or a hedge book that caps the upside.

Rovira explained the choice this way. "It's very difficult for a bank to give you money to go build project A, and before you pay back that debt, you're spending more on project B. That's typically you have early mandatory repayments or cash flow sweeps or cash lock-ups, all those kind of financial covenants ... Just meant it restricted the ability for us to advance Sandstone, which is clearly the flagship asset ... So, how we thought about funding was about maximizing optionality and flexibility."

The cost of that flexibility is 12.5 per cent on US$120 million in a currency the company does not earn. The annual report already records an unrealised foreign exchange loss of A$5.678 million on the bond in the three and a half months to 30 June 2026. Total contractual cash flows on the borrowings are disclosed at A$237.7 million against a carrying amount of A$160.1 million. Rovira intends to refinance once the mill is running, and he says so.

The hedge

Brightstar bought put options over 60,000 ounces of gold in June 2026 at a strike of A$5,809 an ounce, which the company describes as about US$4,000 an ounce equivalent.[R22] A put option is the right to sell at a set price, so it protects against the gold price falling while leaving all of the upside intact. For a company that has just told the market it has no mandatory hedging, it is the right instrument.

The premium is deferred. The cover is published quarter by quarter and it is front-loaded. About 11,100 to 11,400 ounces in each of the four quarters from September 2027, then 6,250, 5,250 and 3,500 in the three quarters to March 2029.[R22] Against the study's 75,000 ounces a year that is roughly 60 per cent of the first full production year covered and exactly 20 per cent of the second. It is a floor under the ramp-up, not a floor under the mine.

Three figures circulate for the cost. The 2 July 2026 announcement calls it "The A$31 million deferred premia settlement".[R22] The accounts then record a derivative asset of A$29.958 million and an option premium payable of the same amount, which is A$499 an ounce. Total contractual cash flows on that premium are A$36.969 million, of which A$23.060 million falls in one to two years and A$13.909 million beyond, which is A$616 an ounce.[R3] Both are the company's figures. The obligation is disclosed clearly, it falls in the same years the bond amortisation is steepest, and it is still an obligation.

The cover also runs out first. The last put settles in the March 2029 quarter. The bond's steepest amortisation and its 20 per cent final payment run through to March 2030, so the last year of repayment carries no floor. Note too that the annual report describes the options two ways, as covering "the first 24 months from FY27" in the directors' report and as settling over the years ending 30 June 2028 and 2029 in the notes.[R3] The 2 July 2026 announcement agrees with the notes. We state both as published.

Three bond terms belong with this and were not stated earlier. The bond carries a 101 per cent change-of-control put, so a takeover triggers a right of repayment at a premium. Early redemption is subject to make-whole payments and call premiums, which is what a refinancing before maturity would have to pay for. And dividends are locked until the project is commissioned, a debt service reserve account is established and the company holds a positive net cash balance after paying them.[R27] That reserve account is a second restricted cash requirement on top of the escrow.

The revenue gap

The Genesis ore purchase agreement completed with the fifth parcel in March 2026. The FY2026 accounts record it as the source of all of that year's revenue, and it was the company's only sales channel in FY2025 as well. Both underground mines, Second Fortune and Fish, are now in care and maintenance, meaning they are shut but kept dewatered, ventilated and able to restart. The fifth and final parcel was processed in the March 2026 quarter and reconciled in the June quarter, and no ore has been processed since. The mill does not commission until the middle of 2027.

So Brightstar has no gold revenue from the March 2026 quarter until mill commissioning, a period of roughly five quarters. Through all of it a 12.5 per cent coupon, which is the annual interest rate the company pays on what it borrowed, accrues on US$120 million and A$101.9 million of capital expenditure commitments run down. The June 2026 quarter shows A$26.95 million of investing outflow and A$18.3 million of that was Laverton mill construction.

This was planned, and it is funded and disclosed. It is neither a surprise nor a solvency question. Total available funding at 30 June 2026 was A$296.7 million. The Appendix 5B is the quarterly cash flow statement every exploration and development company must lodge with the ASX. Item 8.5 of it carries unused finance facilities of A$174.7 million against restricted cash of A$160.9 million in the accounts at the same date, and the company does not reconcile the two.[R3] On the same date the company held ore stockpiles of 130,200 tonnes at 1.74 grams of gold per tonne for 7,261 contained ounces, of which 30,100 tonnes at 3.09 grams a tonne was Second Fortune material.

By 9 September 2026 the company reported Second Fortune stockpiles of about 50,000 tonnes at 3.1 grams for about 5,000 ounces, described as ramp-up material for after plant commissioning. That stockpile is a useful asset for a commissioning campaign.

One number in the quarterly reports should be treated with care. Item 8.7 of the June 2026 Appendix 5B reports "estimated quarters of funding available" as 55.5. That is arithmetically correct and it is economically meaningless, because the calculation divides total available funding by operating outflow only. It excludes the A$26.95 million of investing outflow in the same quarter and the A$101.9 million of committed construction capital. The all-in outflow in that quarter was about A$32.3 million.

For context, at 31 December 2025 the same item read exactly 2.0 quarters[R24], which is the threshold below which the ASX requires a company to answer additional solvency questions. The A$175 million placement was announced 33 days later. The company financed itself out of that position within 33 days, at a 9.5 per cent discount to the 20-day volume weighted average price, and that is the relevant fact.

FY2026 in summary

The A$45.294 million of exploration expensed through the profit and loss includes A$14.3 million relating to an uplift in the rehabilitation provision. Exploration is the single largest driver of the loss, and it is a choice rather than an outcome. A company that capitalised all of it would report a much smaller loss and a larger balance sheet, with no difference in cash.

The auditor is KPMG, appointed during FY2026 in place of Pitcher Partners. The opinion is unqualified with no emphasis of matter and no material uncertainty regarding going concern. Audit fees were A$180,000.

The directors set out the going concern basis in Note 2, and the four conditions they list are the honest summary of the investment case.

"Release of restricted cash which is subject to the completion of certain conditions precedent and continued covenant compliance under the terms of the Nordic bond arrangement. Successful commissioning of the Laverton Mill on time and on budget. Ramp up of operations to achieve acceptable levels of commercial production at budgeted costs. And realising sales within budgeted timeframes at forecast gold prices." Brightstar Resources FY2026 Annual Report, Note 2, 4 September 2026[R3]

Those four conditions are the four things that have to go right. The directors say that if they do not, "the Group may be required to source additional funds through debt or equity markets or a combination of the two". The wording is standard and it is correct. It is also the plainest statement in the annual report of what an investor is underwriting.

16 References and Sources

Every source used in this workup is listed below with its date. Each entry carries an identifier, and the same identifier appears in the body at the point that source is used, so any number, quotation or finding here can be traced back to the document behind it in one step.

[R1] Brightstar Resources Limited, "Updated Goldfields Feasibility Study", ASX announcement, 29 January 2026.

[R2] Brightstar Resources Limited, "Sandstone Mineral Resource Grows to 2.9Moz", ASX announcement, 15 July 2026.

[R3] Brightstar Resources Limited, Annual Report to Shareholders for the year ended 30 June 2026, ASX announcement, 4 September 2026.

[R4] Brightstar Resources Limited, "Menzies and Laverton Gold Projects Feasibility Study", ASX announcement, 30 June 2025.

[R5] Brightstar Resources Limited, "Strategic Acquisition of Aurumin Consolidates Sandstone", ASX announcement, 21 July 2025.

[R6] Brightstar Resources Limited, "Aurumin Scheme Approved - Group MRE Grows to 3.9Moz Au", ASX announcement, 20 November 2025.

[R7] Brightstar Resources Limited, "Group Resource Update Underpins Future Mining", ASX announcement, 19 May 2025.

[R8] Brightstar Resources Limited, "Menzies Mineral Resource increases 22% to 0.7Moz @ 1.5g/t", ASX announcement, 11 December 2025.

[R9] Brightstar Resources Limited, Diggers and Dealers Presentation, ASX announcement, 5 August 2026.

[R10] Brightstar Resources Limited, "Maiden Underground Ore Reserves Underpins FY26 Production", ASX announcement, 26 June 2025.

[R11] Brightstar Resources Limited, "Sandstone Strategic Plan to Deliver Long-Life Production Hub", ASX announcement, 27 January 2026.

[R12] Brightstar Resources Limited, "Lord Byron RC Drilling Results and Mineral Resource Upgrade", ASX announcement, 12 January 2026.

[R13] Brightstar Resources Limited, "Spectacular 305m @ 1.8g/t Au in Sandstone extension drilling", ASX announcement, 3 August 2026.

[R14] Brightstar Resources Limited, "Wide, high-grade intercepts from Two Mile Hill drilling", ASX announcement, 27 August 2026.

[R15] Brightstar Resources Limited, "Significant visible gold in 500m mineralised intersection", ASX announcement, 25 August 2026.

[R16] Brightstar Resources Limited, "High Grade Exploration Success Continues at Sandstone", ASX announcement, 7 September 2026.

[R17] Brightstar Resources Limited, "Sandstone Drilling Reinforces District-Scale Opportunity", ASX announcement, 8 July 2026.

[R18] Brightstar Resources Limited, "EPC Contractor Engaged for Goldfields Project", ASX announcement, 17 February 2026.

[R19] GR Engineering Services Limited, "FY26 Results, Guidance and Equity Raising", ASX announcement, 24 August 2026.

[R20] Brightstar Resources Limited, "BTR Approves Final Investment Decision at Goldfields Project", ASX announcement, 26 May 2026.

[R21] Brightstar Resources Limited, "Laverton Mill Construction Advancing on Schedule", ASX announcement, 9 September 2026.

[R22] Brightstar Resources Limited, "Goldfields Project Development Update", ASX announcement, 2 July 2026.

[R23] Brightstar Resources Limited, "Record Processing Campaign Delivers 7,900oz Au Production", ASX announcement, 20 April 2026.

[R24] Brightstar Resources Limited, "Quarterly Activities Report and Appendix 5B" for the quarter ended 31 December 2025, ASX announcement, 30 January 2026.

[R25] Brightstar Resources Limited, "Quarterly Activities Report and Appendix 5B" for the quarter ended 30 June 2026, ASX announcement, 30 July 2026.

[R26] Brightstar Resources Limited, "Strategic $180M capital raising funds Goldfields development", ASX announcement, 2 February 2026.

[R27] Brightstar Resources Limited, "Brightstar Secures US$120M Bond to Fund Goldfields Project", ASX announcement, 4 March 2026.

[R28] Brightstar Resources Limited, "Funding Package Completed - Fully Funded for Production", ASX announcement, 20 March 2026.

[R29] Brightstar Resources Limited, "Clarification of production targets and aspirational statements", ASX announcement, 14 April 2026.

[R30] Brightstar Resources Limited, "Building Mines and Growing Resources Presentation", ASX announcement, 16 July 2026, and Diggers and Dealers Presentation, 5 August 2026, being the two presentations carrying the Aspirational Statements section.

[R31] Australian Securities Exchange, last traded price, 52-week range and issued capital for Brightstar Resources Limited (BTR), 18 September 2026. The share count is cross-checked against the two Appendix 2A forms Brightstar lodged on 16 September 2026.

[R32] Peer market data, 18 September 2026. Last traded prices from the Australian Securities Exchange for Astral Resources (AAR), Pantoro Gold (PNR), Black Cat Syndicate (BC8), Minerals 260 (MI6), Meeka Metals (MEK) and Ora Banda Mining (OBM). Shares on issue, cash, borrowings, Mineral Resources and Ore Reserves taken from each company's own most recent ASX filing, each dated in the table. Market capitalisation is shares on issue multiplied by the last traded price, calculated by Samso, because several data vendors carry stale share counts.

[R33] gold-api.com for the US dollar gold price and Frankfurter and er-api.com for the AUD/USD exchange rate, all as at 17 to 18 September 2026. The two exchange rate sources agree to three decimal places.

[R34] Interviews and conference presentations given by Alex Rovira, Managing Director of Brightstar Resources Limited, published on YouTube between September 2024 and August 2026. Quotations are taken from the published recordings and every company, project and place name inside a quotation has been confirmed against a primary ASX document. Repeated words and verbal fillers have been removed from the machine-generated captions without ellipses, which is this publication's convention for spoken-word sources.

[R35] Brightstar Resources Limited, "Acquisition of Laverton Tenements Enables Expansion Optionality for Laverton Processing Hub", ASX announcement, 15 September 2026.

[R36] Panther Metals Limited, "Panther Advances Laverton Gold Development Pathway", ASX announcement, 15 September 2026.

[R37] Peer cost track record, Table 22. Study forecasts are taken from six announcements. Meeka Metals, "Expanded Murchison Gold Project Feasibility Study", 12 December 2024. Ora Banda Mining, "Ora Banda Completes Davyhurst DFS", 30 June 2020. Pantoro, "DFS Confirms Attractive Economics and Mine Life for Phase One Norseman Restart", 12 October 2020. Black Cat Syndicate, "Kal East to Generate Strong Returns", 9 May 2024. Astral Resources, "Mandilla Project Pre-Feasibility Study", 25 June 2025. Minerals 260, "PFS confirms Bullabulling's potential to be a High-Margin, Large-Scale, Long-Life Gold Mine", 8 July 2026. Achieved costs are taken from each company's own quarterly activities reports. Those are Meeka 23 October 2025 and 22 July 2026, Ora Banda 29 July 2025 and 16 July 2026, and Pantoro 22 January 2025, 27 October 2025, 22 January 2026, 28 April 2026 and 30 July 2026. Ora Banda's attribution of cost pressure to diesel and industry wide inflation is taken from its FY2026 Annual Report. Black Cat states in its quarterlies of 24 July 2025, 29 January 2026 and 30 July 2026 that it does not calculate an all-in sustaining cost. Its quarterly of 23 April 2026 says the opposite, that guidance including an all-in sustaining cost is expected for FY2027.

17 Disclaimer

The information contained on this website is the writer's personal opinion and is provided to you for information only and is not intended to or nor will it create/induce the creation of any binding legal relations. There may be a conflict of interest present with commercial arrangements with companies and/or stock held. Samso or an associate may receive a commission for funds raised. 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. Accordingly, no warranty whatsoever is given, and no liability whatsoever is accepted for any loss arising, whether directly or indirectly as a result of this information. Investments are subject to investment risks, including possible loss of the principal amount invested. The value of the product and the income from them may fall as well as rise. You may wish to seek advice from an independent financial adviser before committing to purchase or invest in the investment product(s) mentioned herein. If you choose not to do so, you should consider whether the investment product(s) mentioned herein are suitable for you. The writer will not, in any event, be liable to you for any direct/indirect or any other damages of any kind arising from or in connection with your reliance on any information in and/or materials appended herein. The information and/or materials are provided "as is" without warranty of any kind, either express or implied. In particular, no warranty regarding accuracy or fitness for a purpose is given in connection with such information and materials.

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.

The Samso Philosophy:

Stay curious. Stay sharp. And remember, digging deeper always uncovers the real value.

Happy Investing, and the only four-letter word you need to know is DYOR.

Disclaimer

This Samso Research Workup is independent research commentary for general information only. Nothing in this document is financial product advice, and it does not take into account any reader's objectives, financial situation or needs. Figures are drawn from public sources believed reliable at the stated dates but are not guaranteed; market-sensitive numbers change daily. Samso or associated parties may hold positions in, or have commercial arrangements with, companies mentioned. Seek professional advice before making investment decisions.

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.

Samso Research | www.samso.com.au | An Investor Lens on ASX-Listed Companies