Brazil mines 0.5% of rare earths. Its factories lag.

Brazil still holds one of the world's largest reported rare earth reserve positions after the USGS correction. Several deposit types spread across a continent-sized country support that reserve position. USGS still estimated national mine output at 2,000 tonnes during 2025. National output represented about half of one per cent of the world total, a thin result for one of the largest reserve holders. The gap between inventory and flow starts the story while leaving the geological inventory relevant. 1

Brazil's national reserve figure is unusually easy to misuse across project comparisons. ANM's 11.4 million tonnes refers to proven and probable contained oxides in its 2024 inventory. Project announcements often lead with gross tonnes of clay or hard rock at parts per million grades. Those numbers answer different questions and should not be added. 2

ANM recorded R$89.9 million in rare earth exploration expenditure across its national summary. The agency also issued 1,370 new research authorisations across Brazil. Those figures track where money is looking, while actual mine building remains a much later test. Many applications will fail because the ore grade or processing tests disappoint. Others will stall on land access or economics long before a production decision. 2

Pela Ema and Carina are built around weathered, ionic-adsorption-style clays; Caldeira and Colossus follow the same geological model. Their grades look modest beside hard-rock deposits, but simple mining and ambient pressure leaching can reduce crushing and grinding requirements. The strategic appeal lies in the basket: neodymium and praseodymium for high-strength magnets, with dysprosium and terbium that help magnets retain performance at high temperatures. 3

Low grades can still create a large surface footprint because clay projects move substantial material for each tonne of saleable product. Any serious comparison should place water balance and reagent recovery beside recovery rates, then test residue chemistry and restoration across the full mine life. Brazil's official strategy work recognises these execution and environmental constraints. 3

Brazil also hosts carbonatites and alkaline complexes, while mineral sands or existing mine waste could provide other feeds. Vale disclosed a study of rare earth extraction from gold mining residues on 26 August 2026. The study is a useful research avenue without yet supporting a production case, so it stays outside this supply forecast. 14

Serra Verde declared commercial production in January 2024 and originally aimed for at least 5,000 tonnes a year of mixed rare earth carbonate when fully operating. The plant uses heap-style leaching and produces a mixed material containing Nd and Pr. Dy and Tb are also present in the mixed product sold by the mine. A long mine life improves its position, as do existing roads and available power in Goias. Local labour helps as well, but the output still needs separation abroad. 4

USA Rare Earth agreed in April 2026 to acquire Serra Verde in a transaction valued at about US$2.8 billion. The associated plan targets Phase I design capacity of 6,400 tonnes a year by the end of 2027. A US-backed vehicle also arranged a long-term purchase agreement and price support, turning Brazilian raw material into part of an American industrial policy. The acquisition was still conditional at the data cut-off, so the buyer and expansion schedule remain pending. 5 6

Price floors can make non-Chinese supply financeable when spot markets do not. They can also hide the gap between strategic value and ordinary market economics. Sustained production ready for customers at consistent recovery carries more evidential weight than the transaction's headline value.

Aclara filed the Carina feasibility study in April 2026 after opening a pilot plant. Its published case targets about 4,378 tonnes a year of mixed carbonate. The feasibility model assumes an 18-year operating life for Carina. Company scheduling points to 2028, although environmental review and financing still precede any construction. Only a completed build can turn those approvals into consistent first product. 7

Meteoric's Caldeira pre-feasibility study describes a much larger production case of roughly 13,584 tonnes a year. A July 2026 cooperation agreement with POSCO adds a potential customer and financing route, but it is not a final investment decision. The scale makes Caldeira strategically important and puts more pressure on its water case. Product consistency and capital remain separate tests before any final investment decision. 8 9

Viridis proposes about 9,500 tonnes a year from Colossus and has operated a demonstration centre. Management has spoken of Western buyers and a start by the end of 2028. Commercial interest helps, though a demonstration campaign cannot establish the operating record of a full mine and plant. 10

Brazil's June 2026 strategy study is candid about the missing steps. Universities and laboratories give the country a research base, with small technology programmes adding early technical work. At national scale, commercial individual-oxide separation is still missing from Brazil. Brazil also lacks commercial metal-making and sintered-magnet plants capable of industrial output. Building such plants requires technical licences and exacting process control from the first run. Buyers must then qualify the product over long periods before placing large orders. 3

Public financing can narrow part of Brazil's processing and factory technology and construction gap. BNDES and Finep selected 56 strategic mineral plans for possible support. Only a fraction of this multi-mineral pipeline concerns rare earths, which makes the total poor evidence for sector funding. Its R$45.8 billion headline represents proposed investment rather than committed spending. MagBras is a useful magnet technology programme, but its progress should be measured in tonnes sold to customers rather than ministerial visits. 13 12

Permits and public legitimacy will shape which Brazilian projects enter the construction queue. Reporting in August 2026 found exploration interest overlapping or approaching protected land and Indigenous territories. Quilombola communities face similar proximity questions as exploration interest spreads. A credible national strategy needs transparent land screening and real consultation before licensing accelerates. Otherwise the rush for a Western supply label can reproduce the costs it claims to avoid. 15