A Goiás drill hole returned rare earths, niobium and tantalum. The same province's niobium complex has been Chinese-owned since 2016.

On Wednesday, September 23, Rare Earths Americas reported initial drill results from Homer-A, in what the company describes as Brazil’s Goiás Alkaline province, showing near-surface clay and saprolite-hosted mineralisation enriched in the magnet rare earths neodymium, praseodymium, dysprosium and terbium [MINING.COM, 23 September 2026].

Treat the deposit claim with the caution it deserves. There is no mineral resource estimate. There is no demonstrated economic viability. The assays cover roughly 1,250 metres, which the company puts at 8 percent of a planned 15,000-metre programme, and metallurgical testing is still running to establish whether the rare earths are weakly bound to the clay and recoverable by leaching, the property that defines an ionic adsorption deposit [CNN Brasil, 24 September 2026]. The company targets a first resource estimate in 2027.

So the hole is not the story. The element list is, and so is the map it sits on.

What’s happening

• Homer-A assays cover about 2.15 km², which Rare Earths Americas puts at less than 20 percent of the primary geochemical anomaly it has identified and less than 7 percent of a ring-shaped magnetic anomaly exceeding 35 km² and running more than 6.5 km along its major axis [MINING.COM, 23 September 2026]. Desk arithmetic: 2.15 against 35 is about 6.1 percent, consistent with the company’s figure.
• Multiple intercepts report heavy rare earth oxides above 30 percent of total rare earth oxides, NdPr above 25 percent, and combined dysprosium and terbium above 2 percent [MINING.COM, 23 September 2026]. In hole RC-TGA-0010, CNN Brasil reports 22 metres at 2,298 ppm total rare earth oxides plus scandium, including seven metres at 3,424 ppm, with dysprosium and terbium at 4.1 percent of the total in that richer section [CNN Brasil, 24 September 2026].
• Preliminary results also identified zones enriched in niobium and tantalum tied to specific magmatic phases, and scandium returned notable grades in targeted intercepts. All the mineralised intersections, rare earths, niobium and tantalum alike, sit inside the near-surface soil and saprolite weathering profile [MINING.COM, 23 September 2026].
• The programme is a fully funded $5 million, 15,000-metre campaign started in June. The fleet has gone from two rigs to four, more than 6,500 metres are drilled, and the company expects to finish by year-end [MINING.COM, 23 September 2026].
• Worth noting for what it says about how fast this moved: chief executive Donald Swartz said Homer “was not identified as a material project in our S-1 Registration Statement” at the company’s May IPO [MINING.COM, 23 September 2026]. REA closed the day down 1.3 percent at a $219 million market capitalisation.

Read the province, not the hole

Goiás is where four separate element suites are now in play, under four different owners, at four different stages. Set them side by side.

Producing, and Chinese. On 28 April 2016 Anglo American agreed to sell its Niobium and Phosphates businesses to China Molybdenum Co. Ltd for total cash consideration of $1.5 billion. Anglo’s release places the two businesses in the states of Goiás and São Paulo, describes the niobium business as one mine, three processing facilities, two non-operating mines and two further deposits, and names the Catalão complex, the Ouvidor beneficiation plant and the Boa Vista mine among the assets. Together the two businesses generated $146 million of EBITDA in 2015 [Anglo American, 28 April 2016]. CMOC took over the Chapadão and Boa Vista operations in Goiás under that transaction [Dialogue Earth, 5 April 2019]. The province is not speculative. It has been in commercial niobium production for years, and the operating complex has had a Chinese owner for a decade.

Producing, and American. Serra Verde at Minaçu, elsewhere in Goiás, became Brazil’s first commercial rare earth producer in 2024 and is the only scale producer of neodymium, praseodymium, dysprosium and terbium outside Asia. This desk covered its acquisition by USA Rare Earth on 30 May and its 15-year US offtake on 16 June. Those two facts come from the desk’s own archive rather than from reporting done this week.

Mapped, and stalled. INB’s Pró-Urânio programme would lease mining rights for uranium and associated minerals. Of the four mapped areas, Rio Preto in Goiás is the one that matters here: CNN Brasil reports it concentrates uranium combined with copper, gold, rare earths, tin and tantalum. Amorinópolis, also in Goiás, carries uranium with gold, rare earths and diamonds; Espinharas in Paraíba carries gold and rare earths; Figueira in Paraná carries coal [CNN Brasil, 6 July 2026]. A correction to this desk’s own parked note, which had attributed the tin and tantalum suite to both Goiás sites: only Rio Preto carries them.

Drilled, and unproven. Homer-A, as above.

Four legs, one state, and the ownership question is being settled one deposit at a time rather than by policy. One leg went to China in a single 2016 transaction. One went to a US buyer in 2026. One is a state monopoly still structuring a leasing model. The fourth is an early drill programme run by a company with a $219 million market capitalisation.

The part that is actually new

Niobium, tantalum and scandium turning up inside the weathering profile alongside the rare earths is the genuinely novel detail, and it is the one that carries the most AI-materials weight per column inch.

Niobium’s AI link this desk already sized on 23 September, when CBMM put R$13 billion behind oxide lines it says can serve data centers. Scandium goes into aluminium-scandium alloys. Tantalum goes into capacitors. None of these are rare earths, and all of them are things the AI buildout consumes.

The structural point is about co-location rather than grade. If these elements sit in the same near-surface laterite as the rare earths, they are potential by-product credits against a single stripping and leaching operation rather than four separate mines. That is what makes a district economic when individual grades are modest. It is also exactly the logic behind INB’s Rio Preto block, where the state’s uranium monopoly is bundled with copper, gold, rare earths, tin and tantalum and the whole package is waiting on a leasing structure.

The caution is equally structural. Weathering profiles are thin. The average mineralised thickness reported so far at Homer-A is close to 20 metres [CNN Brasil, 24 September 2026]. By-product credits only materialise if the metallurgy recovers four element groups from one flowsheet, and REA has not yet shown it recovers one.

The gate nobody in Goiás has built

Here is the constraint that makes every leg above a Chinese problem rather than a Brazilian solution.

Dysprosium and terbium are the heavies. They are what Ganzhou actually controls, and separating them is the hard step. This desk reported on 19 September, from the ANM’s own statement, that Brazil has fewer than 50 people capable of separating a rare earth and needs 500 by 2040. A high-heavy clay in Goiás does not change that arithmetic. It raises the value of the material that leaves the country unseparated.

Which is why the ownership pattern matters more than the assay. Brazil is accumulating world-class polymetallic geology in one state while the step that converts it into magnet feed remains unbuilt. The 2016 Catalão transaction is what that looks like when it finishes: a producing complex, an offshore owner, and the value-added step happening somewhere else.

What it means

The Tantalum Sovereignty Index sits at 107.5 and the AI Materials Index at 99.7, with materials at 97 and power at 107.9, as of 26 September 2026. The spread between a sovereignty index near its highs and a materials index below par is the whole trade in one line. Ownership and control are repricing faster than the minerals are.

Homer-A does not change any number in those indices, and will not until there is a resource estimate and a metallurgical result. What it changes is the map. A state that was a niobium story and a single rare earth mine is now a four-element district with a live question about who ends up holding it.

What to watch

• Homer-A metallurgy, before year-end. The ionic adsorption test is the binary. If the rare earths leach weakly from the clay, the project has a cheap flowsheet and the niobium, tantalum and scandium credits become worth modelling. If they do not, this is a conventional hard-rock problem with modest grades.
• The Pró-Urânio leasing structure. INB has been structuring this programme without issuing terms. Watch for an actual tender on Rio Preto, and watch whether the tantalum and tin are inside the leased package or carved out with the uranium.
• Who bids for the next Goiás asset. The 2016 precedent is a Chinese buyer taking a producing complex outright. The 2026 precedent is a US buyer taking Serra Verde into antitrust review. The third data point decides whether the first was an anomaly.

Index values are as published in Tantalum’s index file, updated 26 September 2026.