Australian Rare Earths Limited (ASX: AR3)
Abstract
Australian Rare Earths Limited (ASX: AR3) is two very different propositions on one balance sheet. Koppamurra is an ionic clay rare earths project in South Australia and Victoria, which means the rare earths sit in a soft clay blanket near surface rather than locked in hard rock, with a completed Pre-Feasibility Study, a maiden Ore Reserve of 26 million tonnes at 920 parts per million total rare earth oxide, a post-tax net present value of A$858 million and ore running through a pilot plant at ANSTO in Sydney. Overland is a large South Australian uranium exploration position where, in April and May 2026, a drill program looking for sedimentary uranium instead intersected a niobium and rare earth carbonatite in hole OV167, a rare carbonate-rich igneous rock that supplies most of the world's niobium.
What is proven at Koppamurra is the deposit, the flowsheet and the study. The ore is soft clay that needs no drilling and blasting, the rare earths leach out at about pH 2.2, there is no tailings dam, and initial capital is A$178 million for a 12-year operation producing about 1,860 tonnes of total rare earth oxide a year. Roughly a quarter of the product by weight is in the group of rare earths China placed under export control in April and October 2025. What is not proven is a pilot-scale recovery number, which is what the ANSTO program exists to produce, and the offtake, which remains a non-binding memorandum with Neo Performance Materials.
What is proven at Overland is much less, and the workup is deliberate about saying so. One aircore hole returned 19 metres at 0.61 per cent total rare earth oxide from 86 metres to the end of the hole at 105 metres, with the only reported niobium grades being 7 metres at 0.27 per cent Nb2O5 and, inside it, 1 metre at 0.53 per cent. Pyrochlore, the principal niobium ore mineral worldwide, was identified by petrography and scanning electron microscope. Reprocessed magnetics defined a 10 kilometre corridor, and about 2,000 metres of aircore across five traverses started on 1 September 2026. Set against WA1 Resources' Luni at 1.38 per cent Nb2O5 Indicated and Encounter Resources' Aileron high-grade core at 1.7 per cent, AR3's reported niobium grades are a fraction of both. There is no niobium metallurgical testwork at Overland at all.
The constraint that governs everything is the balance sheet. Cash was A$4.607 million at 30 June 2026, there are no financing facilities, and the company's own Appendix 5B reported 2.10 quarters of funding available on its June quarter run rate. Against that is at least A$178 million of development capital, which AR3's disclaimer states there is no certainty of raising when needed, and which it says may be dilutive. At the 3 September 2026 price of A$0.1225, raising A$178 million in equity alone would mean issuing roughly 1.45 billion shares against 256.9 million on issue.
Section 12 compares AR3 with seven listed rare earth and niobium peers on market capitalisation, stage, commodity, jurisdiction risk and the evidence for an economic discovery. Four findings come out of it. AR3 is the smallest company in the comparison by a factor of more than four. Koppamurra is one of only two assets in it with a JORC Ore Reserve, the Code's formal statement that economic extraction has been demonstrated, the other being Arafura's Nolans, which is in construction. South Australia ranks fourth of 68 jurisdictions on the Fraser Institute's 2025 Investment Attractiveness Index. Victoria, where the other five Koppamurra licences are, ranks 49 of 68. And Overland is the only asset in the comparison with no Mineral Resource at all.
The closest technical analogue is Meteoric Resources, whose Caldeira ionic clay project holds 703 million tonnes at 2,617 parts per million total rare earth oxide, about three and a half times Koppamurra's grade, and trades at about twenty times AR3's market capitalisation. Koppamurra is a low-grade deposit even by ionic clay standards. What it has instead is study maturity, a maiden reserve, government backing and a heavy-weighted basket.
The Samso Call on AR3 is WATCH. Both questions that decide what this company is worth, whether Overland is a deposit and how Koppamurra gets funded, have answers arriving within months rather than years. The Overland assays, the ANSTO pilot recovery, the first bulk mixed rare earth oxide sample expected in late September 2026, the Mining Lease Application and any funding announcement are all live. None of them has landed yet. The call describes attention, not action.
Contents
Snapshot - Australian Rare Earths Limited
Prices below are as at the date in the meta line above, the point this workup was compiled from. This is a living document. The share price and market cap will have moved, up or down, in the time since.
Metric | Value |
Share price | A$0.1225 (ASX, 3 September 2026) |
Market capitalisation | A$30.8m (ASX company data feed, 3 September 2026) |
Shares on issue | 256,864,776 (ASX, 3 September 2026). 256,074,545 at 30 June 2026 |
Unlisted options on issue | 46,485,347 (30 June 2026) |
52-week range | A$0.087 to A$0.41 (ASX, 3 September 2026) |
Cash | A$4.607m (30 June 2026, Appendix 5B) |
Financing facilities | Nil (30 June 2026, Appendix 5B item 7.4) |
Quarters of funding available | 2.10 (30 June 2026, Appendix 5B item 8.7) |
Koppamurra Mineral Resource | 243 Mt at 751 ppm TREO (25 June 2026) |
Koppamurra Ore Reserve | 26 Mt at 920 ppm TREO, Probable (25 June 2026) |
Koppamurra PFS post-tax NPV8 | A$858m, IRR 99%, payback 0.9 years (25 June 2026) |
Koppamurra initial capital | A$178m |
Overland status | One drill hole (OV167), 10 km magnetic corridor, ~2,000 m aircore underway from 1 September 2026 |
Government support | A$5m International Partnerships in Critical Minerals grant |
Offtake | Non-binding MOU with Neo Performance Materials, 50% of stage 1 |
Broker coverage | None identified in the sources reviewed for this workup |
Samso Call | WATCH |
Register, as AR3 reports it in the 25 August 2026 presentation. Directors, management and co-founders approximately 15%, Acorn Capital approximately 7%, Neo Performance Materials approximately 1%, and other holders approximately 77%.
1 Why This Company Warrants a Full Workup
Australian Rare Earths Limited is carrying two very different assets on one small balance sheet. Its market capitalisation on 3 September 2026 was about A$31 million. The Pre-Feasibility Study on one of those assets puts a post-tax net present value of A$858 million on it. Section 13 sets out why that gap is the price of funding still to be raised.
The first asset is Koppamurra, an ionic clay rare earths project in south-eastern South Australia and western Victoria. Ionic clay means the rare earth elements are held loosely on the surface of clay particles instead of locked inside hard minerals, so they can be lifted off by ion exchange, which is another chemical simply displacing them. At Koppamurra that is done with dilute acid at about pH 2.2. The alternative, for a hard rock deposit, is crushing, roasting and dissolving, and that is what makes the processing here simple and cheap.
Koppamurra has a Pre-Feasibility Study, a maiden Ore Reserve, a government grant, and ore in a pilot plant at ANSTO in Sydney right now. That A$858 million net present value is a 100 per cent project figure, and AR3 holds Koppamurra 100 per cent, so all of it is attributable to the company.
The second asset is Overland, a very large uranium exploration position in South Australia where, in April and May 2026, AR3 drilled 22 aircore holes looking for sedimentary uranium and instead intersected something else. Hole OV167 returned 19 metres of rare earth and niobium mineralisation, and petrographic work identified pyrochlore, the mineral that supplies most of the world's niobium.
A carbonatite is a rare igneous rock made up mostly of carbonate minerals, and carbonatites are the world's main source of both niobium and rare earths. Section 8 explains why. Reprocessing of an airborne magnetic survey then defined a 10 kilometre magnetic corridor, and drilling to test it started on 1 September 2026.
Koppamurra is an advanced study asset with a funding problem. Overland is a one-hole discovery with a very long runway and no funding attached to it. A workup that looked at only one of them would be describing a different company.
Niobium has also become a live theme on the ASX. WA1 Resources announced the Luni discovery in the West Arunta in Western Australia in November 2022 and built a company the ASX valued at A$831 million on 3 September 2026. Encounter Resources followed with Aileron.
Samso has written about both. So when a third ASX company reports pyrochlore in a carbonatite, the question is whether the numbers reported so far are in the same league. On the public record they are not, and section 9 sets out the comparison.
AR3 went to Overland looking for uranium in sediments. It did not find significant uranium in that program. It found a carbonatite instead. That is worth stating plainly, because a discovery made while looking for something else is a common way that mineral systems get found, and it is also a reason the company had no prior geological framework built around it.
It is also a great reminder that WA 1 Resources Limited went to find IOCGs (Iron Oxide Gold Copper) and they found a world class carbonatite.
2 The Company in Brief
AR3 listed on the ASX on 1 July 2021. It holds 31 tenements across South Australia, Victoria and Queensland, all 100 per cent owned except two Overland licences noted as subject to transfer or ministerial consent. The tenement schedule is in the June 2026 quarterly.
The tenement schedule names nine projects. Three of them carry the story.
Koppamurra is the flagship. Eleven exploration licences, six in South Australia and five in Victoria, about 300 kilometres south-east of Adelaide, within 8 kilometres of a main highway, and about 380 kilometres from the Port of Adelaide. FIG. 1 shows the licence outlines and the towns.
It holds a Mineral Resource of 243 million tonnes at 751 parts per million total rare earth oxide and a maiden Ore Reserve of 26 million tonnes at 920 parts per million. Parts per million is the standard way to report a low-grade, high-volume deposit. 920 parts per million is 0.092 per cent, which sounds trivial until you note that the material is free-digging clay and the rare earths wash out with dilute acid.
Figure 1. The Koppamurra Project, South Australia and Victoria. AR3's location map. The deposit is about 300 kilometres south-east of Adelaide and straddles the South Australian and Victorian border, which is the split section 12 returns to. Naracoorte, Keith, Bordertown and Penola are the towns the South Australian licences are named for. Source: AR3 ASX release, Quarterly Activities Report, 31 July 2026.
Overland is the uranium and now niobium position. It is in South Australia and is described by the company as approximately 8,000 square kilometres in the June 2026 quarterly and the 11 August 2026 release, and as approximately 7,000 square kilometres in the covering note to the 25 August 2026 Tribeca presentation. Both figures are as published by the company. Samso reports them as they stand and does not attempt to reconcile them.
Northern Queensland is nine rare earth exploration permits, grouped under five project names, on which no work was done during the June 2026 quarter.
Two further South Australian uranium licences are outside those three groups. EL7010 at Wirrealpa is the Triggs Bore project and EL7011 at Moolawatana is the Hamilton Creek project. Neither featured in the company's reporting for the period reviewed here. One detail in the schedule is worth recording. EL7079 at Fisher is the only tenement whose listed commodity includes niobium, and it is the licence the carbonatite discovery is in.
The company describes itself as a diversified critical minerals company. On the evidence of where the money went, that is accurate but weighted. A$1.74 million was spent on exploration and evaluation in the June 2026 quarter, and the Managing Director's commentary and the presentation both put Koppamurra first.
3 The Koppamurra Resource and the Reserve
Koppamurra's estimates were updated on 25 June 2026, alongside the Pre-Feasibility Study. Three separate things were reported that day and they are routinely blurred.
A Mineral Resource is mineralisation that has been drilled and estimated well enough to have reasonable prospects of eventual economic extraction. It comes in three confidence categories. Measured is the highest, then Indicated, then Inferred. An Ore Reserve is the part of a Mineral Resource that a study has shown can actually be mined and processed at a profit, after applying mining, processing, cost, price, permitting and environmental assumptions. Only Measured and Indicated material can convert into an Ore Reserve. Inferred material cannot, which is why the reserve is so much smaller than the resource.
An Exploration Target is neither. It is a stated range of possible tonnes and grade, based on insufficient drilling to estimate a resource, and the JORC Code requires it to carry that warning every time it is quoted. The JORC Code is the Australasian Joint Ore Reserves Committee code, the rulebook governing how an Australian listed company may report exploration results, Mineral Resources and Ore Reserves.
More than half the resource is Inferred. 129.6 million tonnes, or about 53 per cent, and the company states that approximately 47 per cent is Measured or Indicated. Inferred material carries low geological confidence and cannot be used in an Ore Reserve.
The grade barely moves between confidence categories. Measured is 773 parts per million, Inferred is 737. That is what an ionic clay blanket looks like. It is laterally continuous, and grade is not the variable drilling resolves.
The resource grew by about 7 million tonnes on the September 2024 estimate with no new drilling at all. The company attributes that to improved geological modelling and re-domaining to isolate calcium-rich material. That is a legitimate technical exercise and it is disclosed as such. It is also not the same thing as finding more rock.
Table 2 gives the reserve. Its grade of 920 parts per million is about 22 per cent above the global resource average of 751. That is the company's own framing and it is arithmetically correct. It reflects the fact that a reserve is a selected, higher-grade, study-tested subset. The deposit did not get better.
Two things behind that uplift are not disclosed. There is no cut-off grade in the quarterly or the presentation, and a cut-off is what turns a resource into a reserve in a deposit where grade barely varies. And the conversion is modest. The 26 million tonne reserve was drawn from the Measured and Indicated portion of a 243 million tonne resource, and the company states that portion is about 47 per cent of the whole. Most of the higher-confidence material did not make the mine plan. That is normal for selective mining of a low-grade blanket and it is why the reserve grade is higher, but the reader should be shown it.
The Exploration Target is a separate figure again. AR3 reports a 2026 Exploration Target of 680 million to 3,620 million tonnes at 540 to 820 parts per million TREO. The JORC Code requires the accompanying statement, which AR3 gives, that the quantity and grade are conceptual, that there has been insufficient exploration to estimate a Mineral Resource, and that it is uncertain whether further exploration will result in one. The range is wide enough, more than five times from bottom to top, that it describes the size of the ground position. It does not describe the size of a deposit. Under the JORC Code an Exploration Target is reported separately from Mineral Resources, so this range is in addition to the 243 million tonne resource and not inclusive of it.
4 The Koppamurra Pre-Feasibility Study Economics
The Pre-Feasibility Study released on 25 June 2026 is the single most important document AR3 has produced. A Pre-Feasibility Study is the second of the three standard study levels. Scoping is first and is indicative. Pre-Feasibility is a preliminary technical and economic assessment good enough to support an Ore Reserve. Definitive Feasibility is the study a board and a lender rely on to commit capital. AR3 is targeting a Definitive Feasibility Study in 2027.
Table 3 sets out the base case as the company reports it. Those are strong numbers and they should be treated as such. A 99 per cent internal rate of return and a payback under a year on A$178 million of capital is unusual. The reason it is unusual is structural rather than promotional. The ore is soft clay that needs no drilling and blasting. There is no tailings dam. The flowsheet is a heap leach at about pH 2.2, which means dilute acid trickled through stacked ore, followed by oxalic acid precipitation to make a mixed rare earth oxide.
Capital that would be spent on crushing, grinding, tanks and tailings storage in a hard rock rare earth project is not spent here. FIG. 2 sets the flowsheet out step by step and FIG. 3 is the company's rendering of the layout.
Figure 2. The Koppamurra process flowsheet. Read left to right. Shallow mining feeds a run-of-mine pad, the ore is agglomerated, then irrigated on a heap leach pad at about pH 2.2. The pregnant leach solution goes to crude mixed rare earth oxide precipitation, re-dissolution and a second precipitation, then to calcination, which the study places offsite and therefore outside the A$178 million capital estimate. Barren material is neutralised and returned to the mined pit. The absence of crushing, grinding and tailings storage is the point of the diagram. Source: AR3 ASX release, Quarterly Activities Report, 31 July 2026.
Figure 3. The company's model of the proposed Koppamurra operation. This is a rendering produced by AR3, not a photograph. Nothing in it is built. It is reproduced because it makes the capital argument visible. The rectangular pads are heap leach cells and there is no tailings storage facility anywhere in the layout, which is the single largest reason the initial capital estimate is A$178 million rather than a multiple of it. Source: AR3 ASX release, Quarterly Activities Report, 31 July 2026.
There is a counterweight to the no-tailings-dam argument and it is geological. AR3's baseline environmental studies, as listed in the quarterly, cover ecology, groundwater, stygofauna and sinkholes. Stygofauna are animals that live in groundwater. Sinkholes and stygofauna together mean karst, which is limestone that groundwater has dissolved into a connected network of voids. The proposal is to irrigate stacked ore with dilute acid above that system, in a district of irrigated agriculture, drawing process water from the local aquifer.
AR3 has done the baseline work and the South Australian Department for Energy and Mining has published a Scoping Report confirming the pathway, so this is already known and already being addressed. But removing the tailings dam removes one class of environmental risk and not this one, and the groundwater case is the part of the Mining Lease Application most likely to draw scrutiny from the people farming the same water.
The Life of Mine Production Target is approximately 71 per cent Probable Ore Reserves and approximately 29 per cent Inferred Mineral Resources. Inferred material has low geological confidence and there is no certainty it converts. The company states that project economics modelled on Ore Reserves alone, with no Inferred material, still return a post-tax NPV8 of A$652 million. That disclosure is exactly what the ASX requires and it is the more conservative number a careful reader should anchor on.
The arithmetic that disclosure implies is worth doing. Twelve years at 3 million tonnes a year is 36 million tonnes of feed. The Ore Reserve is 26 million tonnes. That multiplication is Samso's own, on the company's published figures, and it is why the production target needs Inferred material at all. The reserve on its own does not fill a twelve-year plan.
The presentation goes further and says where the Inferred material is. It is allocated across the whole mine life, at an average of 23 per cent of feed across the first ten years and ranging from 2 to 33 per cent. So the Inferred is not parked at the back of the schedule where drilling could convert it before it is needed. It is in the plan from the early years, which is the more demanding version of the same disclosure and is why the A$652 million reserves-only figure carries weight.
Pricing in the base case uses the average of two independent forecasters, Adamas Intelligence and Argus Media, using Q2 2026 long-term Western forecasts to 2040. Long-term forecast pricing to 2040 is a reasonable basis for a study. It is not a spot price, and rare earth prices in 2026 have been volatile in both directions. The company also models a price floor scenario, which returns a post-tax NPV8 of A$635 million and an 83 per cent IRR.
Capital is broken out as a processing plant at A$77 million, site infrastructure and services at A$48 million, temporary facilities at A$4 million and owner's costs at A$14 million. Those four are the base estimate and they sum to A$143 million. A contingency of roughly A$35 million, about 25 per cent of that base, takes the figure to the A$178 million headline. A 25 per cent contingency is normal at Pre-Feasibility level and is not a warning sign, but it does mean a quarter of the capital number is an allowance rather than a priced item. Operating cost is estimated at A$96 million per year, with mining about 45 per cent and processing about 51 per cent of the total.
Figure 4. Where the A$178 million of Koppamurra initial capital goes. A Samso chart built from AR3's own breakdown. The four base estimate items sum to A$143 million. The contingency of about A$35 million, roughly 25 per cent, sits on top of that figure rather than inside it, so a quarter of the headline capital number is an allowance rather than a priced item. Source: AR3 Pre-Feasibility Study of 25 June 2026, as reported in the Quarterly Activities Report, 31 July 2026.
A second number a reader would look for and will not find is the price side. The study gives an all-in sustaining cost of US$38.32 per kilogram of total rare earth oxide and does not, in the material reviewed here, give the realised price per kilogram behind the revenue. Without it the margin cannot be seen, only inferred from the size of the net present value.
That matters more than usual because the product is a mixed rare earth oxide rather than separated oxides. A mixed oxide is bought at a percentage of the value of the separated oxides it contains, and that percentage, the payability, is set by negotiation with whoever does the separating. What payability the study assumed is not disclosed in the quarterly or the presentation. It is a fair question to put to the company, because a study carrying the full basket value and a study carrying a payability discount describe different projects.
One number a mining engineer would look for is not in the public material reviewed here, and its absence should be stated. There is no strip ratio. A strip ratio is the tonnes of waste that have to be moved to expose one tonne of ore, and it is the single biggest driver of open pit mining cost. Mining is about 45 per cent of the A$96 million annual operating cost, so the strip ratio is embedded in that figure. It is simply not disclosed separately in the quarterly or the presentation. The deposit is described as a shallow blanket in free-digging clay, which implies a low ratio, but implication is not disclosure.
The single hardest sentence in the whole disclosure pack is in the presentation's own disclaimer. To achieve the development outcomes described, AR3 estimates that funding of at least A$178 million will be required, there is no certainty it can be raised when needed, and any such funding may be dilutive. Section 13 deals with what that means against a A$4.6 million cash balance.
Table 4 shows nearly half the basket by weight is cerium and lanthanum, which are the two cheapest rare earths and are in surplus globally. That is true of almost every rare earth deposit and it is not a criticism of Koppamurra. What matters is the other half. Neodymium and praseodymium together are about 21.9 per cent of the calcined product, and dysprosium and terbium together are about 3.2 per cent. Those four are the magnet rare earths that carry the value.
5 Koppamurra Metallurgy and the ANSTO Pilot
Metallurgy decides whether an ionic clay rare earth project works, because the whole economic case depends on the rare earths leaching out cheaply. AR3 has taken a step here that most of its peers have not. It is the first industry partner into ANSTO's clay-hosted pilot plant, which puts it ahead on pilot-scale demonstration even where others are ahead on study stage.
On 11 August 2026 the company announced that commissioning of the pilot plant at ANSTO's new critical minerals facility in Sydney was complete, using about 1 tonne of Koppamurra ore, and that approximately 30 tonnes was now being processed. ANSTO is the Australian Nuclear Science and Technology Organisation. The facility was opened in May 2026 by the Minister for Resources, and AR3 is the first industry partner to use the clay-hosted rare earth pilot plant.
The commissioning phase, completed on 31 July 2026, produced 940 kilograms of on-specification agglomerates, loaded the columns without incident, ran irrigation above design rates with no degradation of flow, and returned rare earth recovery in line with expectations. Agglomerating means binding fine clay into small pellets so that acid can percolate through the stack rather than blinding it. Clay that will not percolate is the classic failure mode for a clay heap leach, so irrigation running above design rate with no flow degradation is the result that carries in that release. FIG. 5 shows the columns.
Figure 5. The ANSTO column leach piloting facility, Sydney. The eight 4-metre columns AR3 is running four cycles through, with about 30 tonnes of Koppamurra ore. AR3 is the first industry partner to use the facility. This is the equipment that will produce the first pilot-scale recovery number, the one the Pre-Feasibility Study's 68 per cent has to survive. One caveat travels with it. A 4 metre column is shorter than a commercial heap lift, and percolation, acid consumption and recovery all shift with bed height. The lift height assumed in the study is not disclosed in the material reviewed here, so the pilot will validate the flowsheet without settling how it behaves at full height. Source: AR3 ASX release, Koppamurra Pilot Processing Underway, 11 August 2026.
“Commissioning is complete and Koppamurra ore is now being processed through a continuously operating pilot circuit — a genuine step up from laboratory-scale testwork. We are now generating engineering data and product samples for direct evaluation by potential customers and offtake partners.
With our first bulk Mixed Rare Earth Oxide sample expected in late September 2026, this program is delivering exactly what we need to finalise Koppamurra's process design and drive toward a bankable, investment-ready outcome.”Travis Beinke, Managing Director and CEO, AR3 ASX release, 11 August 2026.
The claim in that quote worth testing is “bankable, investment-ready”. A bankable feasibility study is a document a lender will lend against. AR3 has a Pre-Feasibility Study and has not yet started a Definitive Feasibility Study, which it targets for 2027. The pilot generates engineering data for that study. So the quote describes a direction of travel accurately and a destination that is at least a study and a funding package away. The rest of it is verifiable and correct.
The program is four cycles of eight columns, each 4 metres tall. Liquor from each cycle goes to oxalate precipitation to make a mixed rare earth oxide. The first bulk sample is expected in late September 2026 and heap leach activities are expected to be complete by December 2026.
The headline recovery of 68 per cent is a study number, not a pilot number. It was carried into the Pre-Feasibility Study, and the company describes commissioning recovery as in line with expectations without quoting a pilot figure. No pilot number is public. That is the number to watch and it is the reason the pilot exists.
The study assumes calcination happens offsite. Calcination is the heating step that turns the precipitated oxalate into a saleable oxide. Doing it offsite keeps capital out of the A$178 million, and it introduces a dependency on a third party that is not named in the material reviewed for this workup.
The recovery figure carries a pH but no leach time and no agglomerate size. A heap leach recovery is a function of both. A 68 per cent recovery at 30 days and 68 per cent at 120 days are very different projects, because leach time sets the size of the pad, the working capital tied up in ore under irrigation, and the rate at which the operation can turn over. The pilot is running four cycles through eight columns each 4 metres tall, on Samso's reading of a release that says four cycles of the eight by four metre columns, and the cycle duration has not been published. It should be one of the numbers the pilot reports.
The material under test is not broken down by domain. It is Koppamurra ore and the announcements say no more than that. The resource was re-domained in June 2026 to isolate calcium-rich material, so at least one recognised domain in the deposit behaves differently. Whether the 30 tonnes at ANSTO is the reserve blend, the best material, or something in between is not stated.
A heap leach at pH 2.2 through calcium-rich clay consumes acid, and the reagent is not named. The 11 August release refers only to lixiviant makeup requirements, a lixiviant being the solution used to dissolve the metal out of the ore. The gap is material. If the lixiviant is sulphate based then calcium does not only consume reagent, it precipitates as gypsum inside the stack, and gypsum blinding is the other classic way a clay heap leach fails alongside the percolation problem. If it is chloride based the question takes a different form. Acid consumption and gypsum behaviour are two of the three variables most likely to move an ionic clay operating cost, and the ANSTO columns are the right place to measure both.
Nobody has yet said where the thorium goes. The Ore Reserve as published carries 19 parts per million thorium oxide and 1.4 parts per million uranium oxide. Those are low by the standards of a monazite hard rock deposit, which is a real advantage. The number that decides things is not the head grade though. It is how much thorium reports to the mixed rare earth oxide, because that governs whether the product travels freely and which separation plants will accept it. The company says the oxalic acid precipitation gives high rejection of impurities and does not say which impurities or to what level. The pilot is the right place to answer that, and the answer belongs alongside the recovery number when it comes.
6 What OV167 Returned at Overland
On 8 July 2026 AR3 announced the identification of a niobium and rare earth carbonatite mineral system at the R254 prospect within the Overland Project. It came out of the Sedan uranium drilling program, which ran 22 aircore holes for 1,280 metres across April and May 2026 to test two sedimentary uranium targets, the Western Flank Zone and the Basal Sandy Unit. That program continued to indicate anomalous uranium in the sedimentary sequence, but no significant uranium intersections were identified from it.
Aircore uses a hollow drill rod with blades that scrape the rock and return chips up the inside of the rod on compressed air. It is fast, it is cheap, and it is the standard first-pass tool in covered terrain. What it returns is cuttings, not core. There is no intact piece of rock to log, measure or orient. Structural measurements are not possible, contacts are approximate, and there is always some risk of down-hole contamination. Aircore answers whether there is something here. It does not answer what its geometry is or how much of it there is.
Figure 6. West to east section through the R254 target, with downhole gamma and assays. The most informative single image AR3 has published on Overland. Four holes are plotted, OV166, OV165, OV167 and OV168, looking north. The stratigraphy runs sand and clay, then sand, then limestone, then clay, then basement. The intercepts labelled on the section, 7 m at 0.56 per cent TREO with 0.27 per cent Nb2O5, 3 m at 1.16 per cent TREO and 9 m at 0.46 per cent TREO, are all on OV167 and all fall at or below the basement contact. The neighbouring holes carry the same stratigraphy and no labelled intercepts. Whether they reached basement at all is the question the section does not answer, and an aircore hole that refused in cover has not tested the target beneath it. Source: AR3 ASX release, Quarterly Activities Report, 31 July 2026.
The hole ended in mineralisation because the hole ended. AR3 reports 19 metres from 86 metres to the end of the hole at 105 metres, with the interval remaining open at depth. Both statements are accurate as the controlling denominator is that the drilling conditions stopped the process of drilling. Aircore drilling stops when the bit will not cut, which in our geological jargon, it is what we call ground refusal as the ground simply became too hard for the drilling process. This usually means that we have reached fresh basement, and 105 metres is where OV167 gave out. Open at depth is a great way to say that nobody has yet tested below it. It implies but does not mean anyone has evidence the mineralisation continues.
The current program does not close that question either. AR3's 1 September release says the open end of hole means the true thickness at OV167 is not yet defined, and then says the program is designed to test that open potential along strike. Along strike is not below. Nothing announced tests what is beneath 105 metres at OV167, and aircore is not the tool that could. The first item on the company's list of evidence for scale is the one item these five traverses leave exactly where it is.
The intervals are downhole lengths and the hole's orientation is not reported. AR3 has published the depths and lengths for OV167 but not its dip or azimuth in the material reviewed for this workup, and it has not reported true widths. True width is the thickness of the mineralised body measured perpendicular to it, and it is what a tonnage calculation needs. Downhole length is what the drill measured. The two are the same only when a hole cuts a body at right angles, and one hole cannot establish the body's orientation in any case. Every length in Table 5 should be read as downhole and as an upper bound on true width. FIG. 6 plots OV167 against its three neighbouring holes.
The niobium grade is reported over 7 metres and 1 metre only. The parent 19 metre interval carries a TREO grade but no niobium grade. The only niobium numbers in the public record are 7 metres at 0.27 per cent Nb2O5 and, inside it, 1 metre at 0.53 per cent. A reader who takes 0.53 per cent as the deposit's niobium grade has taken the best metre of one hole. Section 9 puts those numbers against what a carbonatite niobium deposit needs to be.
Pyrochlore was identified, and that is a real finding. Pyrochlore is the principal niobium ore mineral in the world's major niobium deposits. AR3 identified it by petrography, which means examining the material under a microscope, and by scanning electron microscope work on the drill cuttings. That is a genuine mineralogical identification and it is the difference between "there is niobium in the assay" and "there is niobium in the mineral that industry knows how to process". It says nothing about how much there is.
The company's interpretation is modest. AR3 states that OV167 may have intersected a sill or dyke-like expression of a larger, untested carbonatite intrusive system, and that the hole was drilled above a large coherent magnetic anomaly interpreted to possibly reflect iron and titanium oxide accumulation in basement rocks. A sill or a dyke is a thin sheet. The company is saying it may have clipped an offshoot, not the body. The cautionary note in the quarterly says the results are from a single drill hole, there is insufficient data to define a Mineral Resource, and early-stage results may not be representative. That is the correct disclosure and AR3 has given it without being asked. FIG. 7 shows the anomaly and where OV167 cut its eastern edge.
Figure 7. The OV167 discovery area on total magnetic intensity, with the section line. The close view. OV167 is on the eastern edge of a magnetic high within EL7079. The box at lower right carries the reported intersections. This is the coherent magnetic anomaly AR3 interprets as possibly reflecting iron and titanium oxide accumulation in basement rocks, and it is the reason the company reads OV167 as a possible sill or dyke off a larger untested body. Source: AR3 ASX release, Quarterly Activities Report, 31 July 2026, using South Australian regional total magnetic intensity images from SARIG dataset mesac139.
7 The 10 Kilometre Overland Corridor and the Drilling Now Underway
On 23 July 2026 AR3 reported the result of reprocessing a 1999 airborne magnetic survey flown on 80 metre line spacing. The reprocessing used normalised source strength of the total magnetic intensity, a transform that reduces the distortion a magnetic anomaly suffers when the rock carries remanent magnetism, meaning magnetism locked in when it cooled and pointing in a direction other than today's field. The output was a prominent 10 kilometre north to south magnetic feature extending well beyond OV167. FIG. 8 gives the regional magnetic setting.
Figure 8. Regional total magnetic intensity across the Overland tenements. The wider setting, covering EL7079, EL7081 and the Palmer Fault, the same structure Figure 9 labels the Regional Palmer Fault. Drill collars from the Sedan aircore program are shown, including OV165 to OV168 at the discovery. Magnetics respond to magnetite, so a warm colour here is a magnetic rock and not necessarily a carbonatite. Dolerite dykes, mafic intrusions and iron-rich basement all produce the same response. Source: AR3 ASX release, Quarterly Activities Report, 31 July 2026, using South Australian regional total magnetic intensity images from SARIG dataset mesac139.
Within that corridor, about 3 kilometres south of OV167, is historic hole R64. R64 recorded a significant gamma response from 56 metres to 71 metres. Gamma logging measures natural gamma radiation down a hole. In this setting it responds mostly to thorium and uranium, which travel with many rare earth minerals. A gamma response is a useful pathfinder. It is not an assay and it does not measure niobium at all.
Two further points belong with it. R64 bottomed at 71.0 metres, and at OV167 the mineralisation does not begin until 86 metres, below the base of the cover sequence. On that stratigraphy a gamma response at 56 to 71 metres is likely to be within the cover rather than in basement, and a response in the cover cannot be evidence of a carbonatite beneath it.
Overland is also a sedimentary uranium project. The Sedan program was drilled for sedimentary uranium, and the quarterly reports that it continued to indicate anomalous uranium accumulations in the sedimentary sequence. A gamma high in the cover, in ground where the company keeps reporting exactly that, has a competing explanation with nothing to do with the carbonatite. Re-drilling R64 is the right call because it separates the two, and until it is drilled the historic gamma is the weakest of the three pieces of evidence the company puts behind its scale argument.
R64 was a 120 millimetre diameter hole drilled vertically to a total depth of 71.0 metres. The gamma response therefore runs from 56 metres to the bottom of the hole. This is the same situation as OV167. The response is open at depth because the hole stopped, not because anything below it has been tested. AR3's 1 September release describes it as a significant gamma response over a 15 metre interval, which is the same 56 to 71 metre window stated a different way.
Figure 9. The 10 kilometre magnetic corridor and the five planned drill traverses. The target inventory the current program is testing, from reprocessed 1999 airborne magnetic data flown on 80 metre line spacing. OV167 is at the northern end and R64, shown as R064 on the company’s imagery, about 3 kilometres south. The white bars are the planned traverses, five of them per the release, and the red bracket marks the northern 6 kilometres they cover. The colour scale is normalised source strength in nanotesla per metre, not raw magnetic intensity. The Regional Palmer Fault runs alongside the corridor and the strongest response on the image is in the south, both discussed in the text. What the image shows is a magnetic feature. Whether it is one carbonatite body, a dyke swarm, a structural corridor or something else is what the drilling is for. Source: AR3 ASX release, AR3 commences follow-up drilling to test scale of Overland niobium-REE carbonatite discovery, 1 September 2026.
FIG. 9 shows the corridor and the five planned traverses. The current program started on 1 September 2026. It is approximately 2,000 metres of aircore across five traverses, testing priority targets in the northern 6 kilometres of the corridor, including follow-up on R64. FIG. 10 is the rig on the first hole. AR3 states the program is designed to establish whether mineralisation is continuous along strike and to provide the support required to move from a single discovery toward a defined Exploration Target.
Figure 10. The rig starting the first hole of the follow-up program at Overland. Drilling commenced on 1 September 2026. About 2,000 metres of aircore across five traverses. Assays and downhole gamma data will be reported progressively, and AR3 has not given a date. Source: AR3 ASX release, 1 September 2026.
Travis Beinke, AR3's Managing Director and Chief Executive Officer, put it this way in the release announcing the program.
“The commencement of this drilling program marks the next phase of systematic testing of what is emerging as a significant carbonatite system at Overland. Our initial discovery hole, combined with the subsequently identified 10 kilometer magnetic corridor and encouraging historic drilling results, highlights the scale potential of the system.
This program is designed to test multiple targets along strike and improve our understanding of the geometry, continuity, and fertility of the carbonatite mineralisation.”Travis Beinke, Managing Director and CEO, AR3 ASX release, 1 September 2026.
Read that quote for what it claims and what it does not. It claims the system is emerging and that the program will improve understanding of geometry, continuity and fertility. Those are the three things a second round of drilling is for and the word choice is accurate. It does not claim a grade, a thickness, a tonnage or a resource, and it describes scale potential rather than scale. For a company whose share price is near a 52-week low, that is a restrained piece of language and the restraint is worth crediting.
Two things are visible on AR3's corridor image that the text of the release does not raise, and both cut against reading the corridor as a single body. The corridor runs immediately alongside the Regional Palmer Fault, which is consistent with a deep structure controlling where intrusive rock was emplaced and less consistent with one large intrusion there. And the strongest magnetic response on the whole image is in the south, outside the northern 6 kilometres this program is testing. Neither observation is a criticism. Drilling the northern 6 kilometres is a reasonable choice when that is where the discovery hole and the historic gamma response are. It does mean that a null result across these five traverses would not close out the corridor.
The stated goal of this program is not a Mineral Resource. It is not even close to one. It is an Exploration Target, which is the conceptual range one full step below a Mineral Resource. AR3 is being straight about that and a reader should hold the company to its own framing.
Assays and downhole gamma data will be reported progressively. AR3 has not given a date and this workup does not put one on it. Laboratory turnaround on a program this size is measured in weeks to months rather than days, and that is Samso's expectation from general practice, not a company statement.
8 The Science Behind Carbonatites
A carbonatite is an igneous rock, meaning it crystallised from a melt, that is made up of more than half carbonate minerals. That is unusual. Almost every other igneous rock is built from silicate minerals. Carbonatites come from deep in the mantle, they occur in continental rift settings, and they are rare. There are only a few hundred known in the world. They matter far out of proportion to that number because they are the world's dominant source of both niobium and rare earth elements.
The reason they carry those metals is chemical. Niobium and the rare earths are what geochemists call incompatible elements. They do not fit neatly into the crystal structures of common rock-forming minerals, so as a melt cools and crystallises they stay in the residual liquid and concentrate. A carbonatite melt is an efficient collector of them. When it crystallises, the niobium ends up mostly in pyrochlore, a calcium sodium niobium oxide, and the rare earths end up in minerals such as bastnaesite, monazite and apatite.
The standard reference on how carbonatites are explored is Simandl and Paradis, "Carbonatites: related ore deposits, resources, footprint, and exploration methods", Applied Earth Science, volume 127, issue 4, 2018, pages 123 to 152. It is directly relevant to what AR3 is doing, and it supports part of the approach while raising a question about another part.
What it supports is the method. The paper lists airborne radiometric, magnetic and gravity surveys among the primary tools for finding carbonatites. It cites the Oka complex in Quebec, where the gravity anomaly, the residual total magnetic field, the equivalent thorium response and the first vertical derivative of the magnetic field all coincide with the outline of the complex. That is exactly the toolkit AR3 is using, with magnetics having found the corridor and downhole gamma having flagged R64, and the company has said gravity gradiometry is planned. This is textbook carbonatite exploration and there is nothing improvised about it.
What it raises is geometry. The paper describes undeformed carbonatite complexes as having circular, ring or crescent-shaped aeromagnetic and radiometric signatures. That is the classic signature, and it comes from the fact that a carbonatite complex is typically a roughly cylindrical pipe or plug. AR3's target is a 10 kilometre north to south linear corridor. A linear feature is not the classic signature.
There are at least three ways to read that, and an investor should hold all three at once.
The system is dyke and sill dominated rather than a single plug, which is AR3's stated interpretation of OV167. Carbonatite dyke swarms exist and can be extensive. On this reading the linear corridor is real and the intrusive centres are discrete bodies along it, which is what the five traverses are designed to find.
The corridor is a structural corridor, a deep-seated fault or rift-related zone, along which one or more carbonatite bodies were emplaced. On this reading the corridor is a control on mineralisation rather than mineralisation itself, and most of its 10 kilometres will be barren.
Part or all of the magnetic corridor reflects something else entirely. Dolerite dykes, mafic intrusions and iron-rich basement units are all magnetic. Eighty metre line spacing is detailed rather than coarse, so the limitation is not resolution. It is that total magnetic intensity measures magnetism and not rock type, and cannot discriminate a carbonatite from anything else carrying magnetite. The problem also runs the other way. Carbonatite itself is often only weakly magnetic, and the response commonly comes from associated iron oxide phases or from the altered halo around the intrusion, so magnetics can miss a carbonatite as readily as it can flag something that is not one. AR3's text notes the anomaly under OV167 may reflect iron and titanium oxide accumulation in basement rocks. That is consistent with a carbonatite. It is not exclusive to one.
None of these three is disqualifying and the first is entirely plausible. What the corridor gives AR3 is a list of places worth drilling, which is all a magnetic image can ever give. It is not a discovery, and drilling is what turns one into the other. That is why the current program is the piece of news to follow.
One further point from the same literature deserves stating because it cuts in AR3's favour. Simandl and Paradis note that the high density niobium and rare earth bearing minerals commonly present in carbonatites, pyrochlore among them, are easy to concentrate. They also note that most carbonatite-related apatite deposits currently in production were enriched by weathering. Both matter here. Pyrochlore that can be concentrated by gravity and flotation is a processing advantage over a mineral that cannot, and lateritic weathering is the process that turns modest primary niobium grades into the high-grade blankets that the world's producing niobium mines actually mine. Whether Overland has a weathered profile of that kind is unknown. Nothing in the public record addresses it.
9 What Overland Would Have To Become
Overland's reported numbers are a long way from carrying much weight, and this section sets out why. Table 6 puts AR3's reported niobium numbers against the two ASX carbonatite niobium deposits the market already knows.
Figure 11. Reported niobium grades, AR3 against its ASX carbonatite peers. A Samso figure, built from the sourced figures in Table 6. The two orange bars are single drill intercepts from one hole. The four navy bars are Mineral Resource Estimates. These are not equivalent categories of information, and the chart is a comparison of grade only, not of deposits. Sources as for Table 6.
FIG. 11 puts the six numbers on one scale, and the arithmetic is not close. Luni's Indicated grade of 1.38 per cent Nb2O5 is about five times AR3's best 7 metre intercept and about two and a half times AR3's best single metre. Encounter's high-grade core at 1.7 per cent is higher again. AR3's 0.27 per cent over 7 metres is, on the current public record, an anomalous intercept rather than an ore grade.
That is a statement about one hole, and it should be read as one. Grades in a carbonatite system vary enormously over short distances, and the first hole into a system is very rarely the best one. WA1's own first pass into the West Arunta did not begin at 1.38 per cent either. The reason to state it plainly is that the market has a habit of pricing the word "carbonatite" rather than the number attached to it, and a Samso reader is entitled to see the number.
So what would have to happen for Overland's niobium to matter to AR3's valuation? In order.
The current program has to intersect carbonatite away from OV167. Five traverses across the northern 6 kilometres. If the corridor is a structural feature with one small intrusive on it, this is where that shows up.
Grades have to improve materially. Getting from 0.27 per cent to something starting with a 1 is not an incremental step. It usually requires either a different part of the system, meaning closer to an intrusive centre, or a weathered enriched profile.
The depth has to be liveable. The mineralisation at OV167 begins at 86 metres. Whatever is found along the corridor starts beneath that kind of cover, and every tonne carries the cost of removing or going through the cover first. Nothing in the public record says whether a deposit here would be mined open pit or underground, and at this stage nothing could. Depth is a fixed handicap that better grades have to overcome, and it does not improve with drilling.
Thickness has to be established with drilling that can measure it. Aircore that stops at 105 metres cannot define the geometry of a body. Diamond drilling, which recovers an intact cylinder of rock, is what gives true widths, structural measurements and material for proper metallurgical testing.
Metallurgy has to be done, and none has been. There is no niobium metallurgical testwork at Overland in the public record. Pyrochlore recovery from carbonatite is normally a flotation problem. Flotation separates minerals by making the ones you want stick to air bubbles and float off. It is sensitive to the gangue mineralogy, the gangue being the worthless minerals the ore mineral is mixed with, and particularly to fine-grained iron oxides and to the carbonate itself. Both WA1 and Encounter have spent years and considerable money on this step. The identification of pyrochlore under a microscope is the beginning of that work, not a substitute for it.
Someone has to fund it. Section 13.
On any reasonable reading, steps one to five are a multi-year program. That is not a criticism of AR3. It is the normal timeline for a carbonatite discovery.
10 The Niobium Market
Niobium is a small market with an unusual structure, and understanding the structure explains why an ASX discovery gets rerated so hard.
Roughly 90 per cent of niobium goes into steel, as ferroniobium, an iron and niobium alloy. Adding a fraction of a per cent of niobium to steel makes it dramatically stronger, which is why it goes into pipelines, car bodies, structural steel and ships. The remainder goes mostly into superalloys for jet engine turbines, with small volumes in superconducting magnets and electronics. An emerging use is niobium titanium oxide anodes for fast-charging batteries, where CBMM has stated a target of 20,000 tonnes by 2030.
Price is where niobium becomes difficult to model. It does not trade on an exchange and there is no transparent daily benchmark. 85 to 90 per cent of sales are made under long-term contracts, so published prices are indicative rather than clearing prices. The USGS weighted average unit value for ferroniobium in 2024 was US$26 per kilogram on a gross weight basis.
Shanghai Metals Market quoted 60 per cent ferroniobium at US$48,667 per tonne on 3 August 2026 and 99.5 per cent niobium pentoxide at US$56.86 per kilogram, the latter down 7.8 per cent from US$61.70 in July. Samso's own July 2026 work used approximately US$30,000 per tonne for ferroniobium and around US$45,000 per tonne for niobium oxide. Those figures are not directly comparable to each other because they are different products, different purities, different bases and different reporting dates. Anyone modelling niobium revenue needs to be careful about which one they have picked up.
The whole global niobium market is worth roughly US$5 billion a year. For comparison, that is a fraction of the copper or iron ore markets. It is a small, sticky, contract-driven market controlled by one dominant private company. That is precisely why Western governments list niobium as a critical mineral and why a credible new source outside Brazil attracts attention. Niobium is on Australia's Critical Minerals List. Table 7 sets out who produces it.
The counterweight is that a small, contract-driven market is a hard market to enter. There is no spot market to sell into. A new producer has to displace CBMM in someone's supply chain, and CBMM has spent decades making that difficult, on price and on technical service. The strategic argument for a Western niobium source is real. Whether a new entrant can find a commercial path into a contract-bound market is a separate question.
11 The Rare Earth Market
The rare earth market AR3 is aiming at looks very different from the niobium market, and the difference is the reason Koppamurra has a Pre-Feasibility Study while Overland has one hole.
Rare earths are seventeen elements, and only a handful of them carry real value. Neodymium and praseodymium, usually quoted together as NdPr, are the workhorses of permanent magnets. Dysprosium and terbium are added in small quantities to stop those magnets losing their strength when they get hot, which is why they are indispensable in electric vehicle motors, wind turbines and defence hardware. Those four are the magnet rare earths. Cerium and lanthanum are abundant, cheap and in surplus.
Two things have reshaped this market since 2025.
The first is Chinese export control. In April 2025 China placed export controls on seven rare earth elements, samarium, gadolinium, terbium, dysprosium, lutetium, scandium and yttrium. In October 2025 it added five more, holmium, erbium, thulium, europium and ytterbium, under Ministry of Commerce announcement number 61 of 2025, with most measures effective from 1 December 2025. Twelve of the seventeen rare earths are now covered.
Every rare earth in the Koppamurra basket other than lanthanum, cerium, praseodymium and neodymium is in that controlled group. Table 4 shows only the larger oxides, so the controlled share of the product is higher than the five entries visible there suggest. Yttrium alone is 16.2 per cent of the calcined product, with samarium at 3.4, gadolinium at 3.3, dysprosium at 2.7 and terbium at 0.5. AR3 makes this point in its material and the point is fair.
The second is the arrival of price floors. In July 2025 MP Materials and the United States Department of Defense entered a ten-year agreement with a minimum price of US$110 per kilogram for NdPr. In March 2026 Lynas signed a restructured long-term supply agreement with Japan Australia Rare Earths running to 2038, at a minimum of US$110 per kilogram for 5,000 tonnes a year of NdPr oxide, with upside sharing above US$150.
In April 2026 USA Rare Earth agreed to acquire Serra Verde Group for US$2.8 billion, with a fifteen-year full offtake for Pela Ema in Brazil and guaranteed floors of US$110 per kilogram NdPr, US$575 per kilogram dysprosium and US$2,050 per kilogram terbium, backed by United States government agencies. Australia has flagged a Critical Minerals Strategic Reserve expected to be underpinned by national offtake agreements including rare earths.
A floor price changes what a rare earth project is. It converts an unbankable commodity price assumption into something a lender can underwrite. That is why AR3 puts those three deals on a slide, and it is a legitimate thing to put on a slide.
The counterweight is the spot market. NdPr oxide was quoted at US$97.40 per kilogram on 4 August 2026, down about 11.9 per cent from US$110.55 in July, which put it below the US$110 floor level for the first time in that series. Dysprosium oxide was US$182.91 per kilogram and terbium oxide US$878.63, both down on the month, dysprosium by about 12 per cent and terbium by about 7.6 per cent. Those are Shanghai-referenced indicative prices and they are not the price a Western offtake would be struck at. They do show that the physical market in 2026 has been volatile in both directions, and that the price floors are floors precisely because the spot market is not reliably above them. FIG. 12 sets the two against each other.
Figure 12. Where the spot market was against the contracted floors, August 2026. A Samso figure, built from figures already sourced in this section. The floors are the guaranteed minimum prices in the USA Rare Earth and Serra Verde agreement announced in April 2026, backed by United States government agencies. The spot marks are Shanghai-referenced indicative prices as at 4 August 2026. These are not the same market and a floor struck in a Western offtake is not a price a Chinese spot index has to respect, which is the whole reason floors exist. What the chart shows is how far below those floors the physical market was on the day. Floors from the AR3 Tribeca Future Facing Symposium presentation, 25 August 2026. Spot marks from critical-minerals-news.com, 4 August 2026.
Two further points a Koppamurra investor should hold.
AR3 has a non-binding memorandum of understanding with Neo Performance Materials for 50 per cent offtake of stage one. An offtake agreement is a commitment by a buyer to take a set volume of production, usually over a set period. Non-binding is doing real work in that sentence. A memorandum of understanding is a statement of intent. It is not a contract, it carries no price, and it is not finance. It is still meaningful, because Neo is a real processor and it also holds about 1 per cent of the register, but it should not be read as offtake secured.
And the product is a mixed rare earth oxide, not separated oxides. A mixed oxide has to go to a separation plant somewhere, and the separation industry outside China is small. That is a supply chain dependency the Pre-Feasibility Study economics depend on.
12 Peer Comparison
A peer comparison is the fastest way to see what the market is actually paying for, and it only works if the comparison is made on stated grounds. This section compares AR3 with seven other listed rare earth and niobium companies across five dimensions.
Market capitalisation. What the market pays for the whole company today, read from the ASX company data feed on 3 September 2026.
Stage. Where the asset stands on the sequence from a drill intercept, to a Mineral Resource, to a study, to an Ore Reserve, to a final investment decision. This is the dimension that moves valuation most, because each step removes a category of risk.
Commodity. Rare earths, niobium, or both, and within rare earths whether the basket is weighted to the heavy magnet elements that are hardest to source outside China.
Jurisdiction risk. Measured by the Fraser Institute's Investment Attractiveness Index, which surveys mining company executives and scores 68 jurisdictions on geology and on policy together. It is a survey of perception, not a legal opinion, and that is exactly why it is useful here. Perception is what sets the discount rate a market applies.
Geological potential for an economic discovery. The hardest of the five and the one that cannot be reduced to a number. Samso's approach is to state the evidence that exists for each asset and then to place it on a four-step scale, with the scale defined so a reader can disagree with the placement and still use the evidence.
The comparison is not a ranking and it is not a valuation. These companies are at very different stages, and a company with a final investment decision is not competing with a company holding one drill hole. There is no target, no multiple and no implied price for anything in the table.
Market capitalisation
AR3 is the smallest company in Table 8 by a wide margin. At about A$31 million it is roughly one fifth of Encounter Resources, one sixth of Victory Metals, one twentieth of Meteoric, one twenty-seventh of WA1 Resources and roughly one thirty-fifth of Brazilian Rare Earths and Arafura.
Some of that gap is stage and some of it is deposit quality, and this section's job is to separate the two. What the raw number establishes is the starting point. AR3 is priced as a company that has not yet solved its funding problem, and the other seven are priced as companies further along, holding better deposits, or both.
Stage
On stage alone AR3's Koppamurra ranks higher than most of the table.
Arafura is the only company here past a final investment decision. After that the order runs as follows. Koppamurra, which has a Pre-Feasibility Study and a maiden Ore Reserve. Then Caldeira and Luni, which have established resources and studies in progress but no reserve. Then Rocha da Rocha and North Stanmore at scoping study level. Then Aileron with an Inferred resource. Last, Overland with one drill hole.
An Ore Reserve is the step most of this table has not taken. Under the JORC Code an Ore Reserve is the economically mineable part of a Measured or Indicated Mineral Resource, declared only after a study has applied mining, processing, cost, price, marketing, legal, environmental, social and governmental factors. It is the Code's formal statement that somebody has demonstrated economic extraction. Of the eight assets in Table 8, two have one, Nolans and Koppamurra.
Commodity
The table splits into three commodity positions.
Arafura and Brazilian Rare Earths are light-weighted. Their value is mostly in neodymium and praseodymium, the two elements that carry most of the value in a rare earth basket.
Meteoric, Victory Metals and AR3's Koppamurra are the ionic and clay-hosted group, where the attraction is a higher proportion of heavy rare earths, dysprosium, terbium and yttrium among them, and cheap leaching. Victory Metals reports a heavy share of total rare earth oxide averaging 39 per cent. AR3 does not report Koppamurra on that basis in the material reviewed here, so the two cannot be set against each other directly. What AR3 does report is that magnet rare earths are approximately 25 per cent of total rare earth oxide in the Mineral Resource, which counts a different set of elements. Heavy is also not a fixed boundary between companies, so a like for like comparison would need both definitions before it meant anything.
WA1 and Encounter are pure niobium, which is a separate market with separate customers and no overlap in the buyer base at all.
AR3 is the only company in this table with exposure to both commodities, and that is a real differentiator. It is also the reason its treasury has to be split, which section 13 deals with.
Jurisdiction risk
FIG. 13 plots the Investment Attractiveness Index score for every jurisdiction in Table 8.
Figure 13. How mining executives rank the jurisdictions behind each asset in Table 8. A Samso figure, built from the Fraser Institute’s published scores. The index combines a view of geology with a view of policy, and it is a survey of perception rather than a legal assessment. South Australia, where six of the eleven Koppamurra licences and all of Overland are, ranks fourth of 68. Victoria, where the other five Koppamurra licences are, ranks 49 of 68. Source: Fraser Institute, Annual Survey of Mining Companies 2025, Investment Attractiveness Index.
South Australia ranks 4 of 68 on the Fraser Institute's Investment Attractiveness Index for 2025, with a score of 89.19, ahead of Western Australia at 6 with 87.01. Brazil ranks 19 with 76.74. The Northern Territory, where Nolans is, ranks 33 with 68.10.
Victoria ranks 49 of 68, with 59.52. That is the lowest score attached to any asset in this comparison, and it is attached to part of Koppamurra. AR3 does not raise the point in its own material. AR3's eleven Koppamurra licences are split across the border, six in South Australia and five in Victoria, and the deposit is one continuous blanket that does not care about the state line.
What AR3 has said about where development happens first points to South Australia. The Mining Lease Application is proceeding through the South Australian Department for Energy and Mining and the Scoping Report came from the same department. Overland is entirely in South Australia too, in a state that hosts the only three currently operating uranium mines in Australia. So the near-term work is on the high-ranking side of the border and the Victorian ground is longer-dated optionality.
One caution on the index itself. It is a survey of mining executives' perceptions, it moves year to year, and a high score reflects a favourable view of both geology and policy combined. The Fraser Institute also publishes a Policy Perception Index that strips the geology out. On that measure Western Australia ranks 13, South Australia 16 and Victoria 34. The gap between Victoria's policy rank of 34 and its overall rank of 49 says that respondents think less of Victoria's geological endowment than of its policy, which is a different problem from the one usually assumed.
Geological potential for an economic discovery
This dimension cannot be measured and Samso will not pretend otherwise. What can be done is to state the evidence that exists for each asset, place it on a defined scale, and let a reader who disagrees with the placement still use the evidence.
Three things fall out of Table 9.
The first is that Koppamurra is one of only two assets in this comparison carrying a JORC Ore Reserve, the other being Arafura's Nolans, and it carries the smallest market capitalisation in Table 8 by a factor of more than four. An Ore Reserve is the JORC Code's answer to the question this dimension asks, and Koppamurra has one. Meteoric, which does not, trades at about twenty times AR3. That is the strongest argument in AR3's favour anywhere in this workup.
The second is the counterweight, and it is heavy. An Ore Reserve is a statement about rock, not about money. Nolans has a reserve, a final investment decision and A$1.15 billion of market value, and the distance between Koppamurra and Nolans is the distance AR3 has to travel. Grade is also part of the answer. At 751 parts per million globally, Koppamurra is the lowest-grade rare earth deposit in this table by a wide margin, and low grade is what makes the project dependent on the heap leach staying cheap and the recovery staying at 68 per cent. The pilot at ANSTO is testing exactly that.
The third is that Overland is the only asset in this comparison placed at Untested, and it is the asset that has attracted the attention. One hole, no resource, no Exploration Target yet, no metallurgy, and reported niobium grades a fifth of Luni's Indicated grade. Nothing about that makes it a poor prospect. Carbonatite systems are found this way and both WA1 and Encounter started with less certainty than they have now. It does mean that anyone pricing AR3 on the carbonatite is pricing a target inventory, and Table 9 is the reason this workup places the weight where it does.
What the comparison says about AR3
AR3 is the smallest company in Table 8 by market capitalisation. It holds one of only two assets in the table carrying a JORC Ore Reserve. Its near-term development work is in South Australia, the highest-ranked jurisdiction in the table. And it owns the least advanced exploration asset in the table. All four are true at once, and any argument about AR3 that uses only one of them is incomplete.
The gap between AR3 and its closest technical analogue is mostly explained by the deposit rather than by the market. Caldeira is about three and a half times the grade of Koppamurra and its resource is several times larger. Samso has not confirmed how much of Meteoric's 703 million tonnes is Measured and Indicated, so the higher-confidence comparison is not made here. Meteoric trades at about twenty times AR3. Koppamurra is a low-grade deposit even by ionic clay standards. What it has that grade alone does not capture is study maturity, a maiden reserve, a government grant, ore in a pilot plant, and a heavy-weighted basket.
The ratio AR3 puts in front of investors is a different one again. The company shows its Pre-Feasibility Study net present value of A$858 million against a market capitalisation it stated at about A$28 million, a ratio of about 0.03 times. That arithmetic is correct. What it does not price is the A$178 million of capital that has to be raised before any of that value is realised, by a company with A$4.6 million in the bank. A market capitalisation to net present value ratio of 0.03 on an unfunded pre-development project is not a mispricing on its own. It is what the market pays when it expects heavy dilution between here and there. Section 13 puts numbers on that.
13 The Balance Sheet and the Funding Gap
An Appendix 5B is the quarterly cash flow report every ASX exploration company must lodge, on a form the exchange prescribes. Item 8.7 on that form is the company's own calculation of how many quarters its cash will last at the rate it just spent it. Table 10 carries it.
That number decides the rest of this section. On its June quarter run rate, AR3 reported 2.10 quarters of funding available. That is the company's own calculation on the formula the ASX prescribes. It does not include the A$2.1 million grant instalment expected in August 2026, which pushes the position out, and it is a snapshot of a quarter in which A$1.74 million went into exploration and evaluation. But it is the number, and it says AR3 is a company that will need to raise money.
Against that, A$178 million of development capital. At the 3 September 2026 price of A$0.1225, raising A$178 million in equity would require issuing roughly 1.45 billion shares, against 256.9 million on issue. That is not a forecast and it is not a suggestion that the company will do it that way. It is arithmetic, and it is the reason the market capitalisation to net present value ratio is 0.03 rather than 0.3. Real development funding for a project of this type normally comes as a package of debt, strategic equity, offtake-linked finance and government support, and AR3 says it is assessing all of those. The company also states plainly that no binding development funding commitment has been received.
Figure 14. What AR3 has against what Koppamurra needs, on one scale. A Samso chart built from the company's own reported figures. Cash, market capitalisation, the initial capital requirement and the study net present value are drawn on the same axis, which is what makes the distance between them visible. Sources: AR3 Appendix 5B and Quarterly Activities Report, 31 July 2026, ASX company data feed read 3 September 2026, and the Pre-Feasibility Study of 25 June 2026.
The options are a modest offset. Of the 46,485,347 options on issue, 16,374,926 are exercisable at A$0.10 and expire on 30 March 2029. At the current price they are in the money, meaning the exercise price is below the share price so a holder gains by exercising. Full exercise of that tranche would raise about A$1.64 million and issue about 6.4 per cent of the current share count. The next tranche up, 18,189,194 at A$0.2775 expiring 21 April 2029, is well out of the money at A$0.1225.
The balance sheet funds exploration and the pilot program. It does not fund a mine, and it does not fund a serious multi-year carbonatite campaign at Overland at the same time. Every dollar spent at Overland is a dollar not spent advancing Koppamurra, and a company with two assets and one small treasury has to choose. So far the choice has been Koppamurra, and the Overland program at approximately 2,000 metres of aircore is small enough to be affordable alongside it.
14 Management Weighed
The board and executive are four people. The four descriptions below are as AR3 states them on the team slide of its 25 August 2026 presentation. Samso has not independently verified the individual career claims.
Angus Barker, Independent Chairman. Over 30 years of professional experience, including at top-tier global investment banks and as a senior adviser to Australian Government Ministers in key economic portfolios. That second half carries weight for a company whose funding case leans on the A$5 million International Partnerships in Critical Minerals grant reported in the 31 July 2026 quarterly and on a prospective Critical Minerals Strategic Reserve. Government-facing experience is not decoration in this sector in 2026.
Travis Beinke, Managing Director and Chief Executive Officer. Over 20 years with Australian and internationally listed resource companies. He is the public face of the company, presenting at RIU Sydney in May 2026 and at the Tribeca Future Facing Symposium in Singapore in August 2026. His commentary in the announcements reviewed for this workup is measured. The 1 September release quoted in section 7 describes Overland as an emerging carbonatite system with scale potential, and claims nothing about grade, thickness or tonnage. Secondary reporting of the 8 July release has him describing the discovery as an unexpected expansion of the company's critical minerals exposure. Samso has not read the 8 July release directly and reports that characterisation at second hand.
Rick Pobjoy, Chief Technical Officer and Co-Founder. Geologist with more than 25 years in mining and mineral exploration. He is the named Competent Person for exploration results and the Exploration Target. A Competent Person is the individual who takes personal responsibility under the JORC Code for a reported estimate. A co-founder still holding the technical role is a positive on a company this size, because the person compiling the exploration results has been with the ground from the beginning.
Two qualifications belong with that. Mr Pobjoy is named as the Competent Person for exploration results and for the Exploration Target, not for the Mineral Resource or the Ore Reserve, and the Competent Persons for those two estimates are not named in the material reviewed for this workup. Those are the estimates the Pre-Feasibility Study rests on. And Mr Pobjoy is an internal Competent Person, being an officer and co-founder of the company. That is permitted under the JORC Code and is normal at this size, and it stands in contrast to the metallurgy, which is signed by an external Competent Person.
Pauline Carr, Independent Non-executive Director. Experienced company director in compliance, governance and risk across 30 years in the resources sector.
Metallurgical results are signed off by James Davidson, principal metallurgist of Rendement and a Fellow of the Australasian Institute of Mining and Metallurgy, as an external Competent Person.
On the register, the company reports directors, management and co-founders holding about 15 per cent, Acorn Capital about 7 per cent, Neo Performance Materials about 1 per cent, and other holders about 77 per cent. Fifteen per cent held by the people running the company is meaningful alignment at this size. Acorn Capital is a specialist Australian microcap investor, and its presence is a mild positive signal on institutional diligence, with the obvious caveat that a specialist microcap fund holds many positions.
Figure 15. The AR3 register as the company reports it. A Samso chart built from AR3's own presentation. Directors, management and co-founders hold about 15 per cent, Acorn Capital about 7 per cent and Neo Performance Materials about 1 per cent, leaving about 77 per cent with other holders. The percentages are approximate and are the company's own. Source: AR3 Tribeca Future Facing Symposium presentation, 25 August 2026.
The judgement Samso would offer is this. The disclosure quality is above average for a company of this size. The cautionary statements in the 8 July and 31 July releases were volunteered, not extracted. The Overland language has stayed proportionate to a one-hole result through two months of an obvious opportunity to do otherwise. The presentation is a different document with a different job, and its arithmetic is looser in places, which section 15 returns to. On balance, a reader should treat AR3's announcements as reliable and its presentation as marketing, which is the normal and correct way to read both.
15 Samso Concluding Comments
AR3 holds two assets at opposite ends of the development sequence under one ticker, and the Samso Call reflects the one that is further from being resolved.
The Koppamurra half is real, advanced and priced well below the net present value its own study carries. A Pre-Feasibility Study, a maiden Ore Reserve, a government grant, ore running through a pilot plant at ANSTO, a defined regulatory pathway and a Mining Lease Application targeted for 2026 are not nothing. Very few A$31 million companies on the ASX have that stack. The heap leach flowsheet is low capital and low technical risk by rare earth standards, and the product basket is weighted toward the elements the West cannot easily buy.
The Koppamurra half also has one problem that dominates all its virtues. A$178 million of development capital against A$4.6 million of cash, no debt facility, no binding funding commitment and a share price at A$0.1225. Everything good about the project has to survive that gap. The market is not failing to see the A$858 million net present value. It is pricing the dilution required to get there, and on the current share price that dilution would be severe.
The Overland half is interesting and it is very early. Pyrochlore identified by microscope and scanning electron microscope in a carbonatite is a real mineralogical finding. A 10 kilometre magnetic corridor is a real target inventory. Drilling started on 1 September 2026 and results will come. What the public record does not contain is a niobium grade that stands comparison with Luni or Aileron, a second hole, a true width, a thickness, a tonnage, or a single piece of niobium metallurgical testwork. The company's own stated goal for this program is an Exploration Target, which is one step below a Mineral Resource. Nobody should be pricing an Overland niobium deposit today, because there is not yet enough information to describe one.
There is a version of the next twelve months where both halves work. The pilot delivers a recovery number that holds, product samples convert a non-binding memorandum into something binding, the Mining Lease Application goes in, the Definitive Feasibility Study starts, a strategic or government funding package appears, and the Overland traverses hit carbonatite with better grades than OV167. In that version A$31 million looks like a mistake.
There is also a version where the aircore traverses return more anomalous but sub-ore-grade intercepts, the Definitive Feasibility Study needs money the company has to raise at these levels, and the register is diluted well before any of the Koppamurra value is realised. Neither version is remote.
What separates them is information that does not exist yet and will exist within months. That information is the Overland assays, the pilot plant recovery, the first bulk mixed rare earth oxide sample expected in late September 2026, the Mining Lease Application, and any funding announcement at all.
That is why the Samso Call on AR3 is WATCH. The assets are not poor and the company is not doing anything wrong. The reason is that the two questions deciding what this company is worth, whether Overland is a deposit and how Koppamurra gets funded, both have answers arriving and neither has arrived. The call means watch this company rather than act on it. Samso publishes no price targets and no valuations.
Figure 16. What AR3 has said it will deliver, and when. A Samso figure, assembled from dates the company has stated. Sources, in order. The first bulk mixed rare earth oxide sample and heap leach completion from the AR3 ASX release of 11 August 2026. The Overland assay timing is Samso’s note that AR3 has committed to progressive reporting without giving a date, per the ASX release of 1 September 2026. The Mining Lease Application and the Definitive Feasibility Study from the AR3 Quarterly Activities Report of 31 July 2026. Final investment decision, construction and first production from the AR3 Tribeca Future Facing Symposium presentation of 25 August 2026. Every date is the company’s target, not a commitment, and each depends on funding that has not been secured.
FIG. 16 sets the stated dates out in order. Three things would move this call.
Overland assays showing niobium grades materially above 0.27 per cent Nb2O5 over useful widths, in more than one hole, along the corridor. That would turn a discovery into a system and would justify reading Overland as a second asset rather than an option.
A funding structure for Koppamurra that is not straight equity at these levels. Government support, strategic equity from a customer, or offtake-linked debt would all change the dilution arithmetic that currently caps the share price.
A pilot plant recovery number at or above the 68 per cent carried in the Pre-Feasibility Study, with the acid consumption disclosed. That is the single most direct de-risking event available to the company in the next quarter.
Two housekeeping notes for completeness, in line with Samso's practice of reporting what a company has published rather than reconciling it. AR3 describes Overland as approximately 8,000 square kilometres in the 30 June 2026 quarterly and the 11 August 2026 release, and as approximately 7,000 square kilometres in the covering note to the 25 August 2026 presentation. The same 25 August presentation states a market capitalisation of about A$28 million based on a A$0.11 share price on one slide, and A$26.9 million based on a A$0.105 share price as at 19 August 2026 on another. Both sets of figures are as published by the company. Resolving them is the company's business, not Samso's, and neither affects the analysis above.
The question a reader is left holding is a simple one. If the Overland assays and the pilot recovery number both arrive inside the next two quarters, which of the two actually changes what AR3 is worth?
16 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.
 The Samso Way – Seek the ResearchHere 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:
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. DisclaimerDisclosure of interest. Samso Research Media House holds a small shareholding in Ordell Minerals Limited (ASX: ORD). This workup was not commissioned by Ordell Minerals and Samso received no payment from the company for producing it. 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. Seek professional advice before making investment decisions. About SamsoSamso 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 |
Comments