Follow the Traders: The Hidden Power Behind the New Rare Earth Order-At Least for Now

Highlights

• China's export controls pushed European gallium and heavy rare earth prices to roughly five times Chinese domestic prices, triggering urgent supply chain restructuring by Western traders and governments.
• Traxys, Mercuria, and Hartree Partners emerged as inaugural suppliers to the U.S. EXIM Project Vault strategic reserve, signaling traders' new role as critical infrastructure rather than passive middlemen.
• Japanese trading houses Sojitz and Sumitomo aggressively locked long-term offtakes with Lynas and Ucore, extending supply commitments to 2038 and embedding price floors to secure rare earth corridors for Japan.
• POSCO International moved beyond procurement to build a full rare-earth value chain, launching a corporate venture fund and a U.S. joint venture targeting 6,000 tonnes of separation and magnet production capacity.
• Despite headline deal volume, many H1 2026 arrangements remain non-binding term sheets or letters of intent, underscoring that traders are buying time and optionality rather than solving the West's magnet supply gap.

---

In the first half of 2026, rare earth, critical mineral, and specialty-metals traders stopped behaving like quiet middlemen and started to behave, at least in some cases, like strategic infrastructure. The reason was simple: the underlying market remains extraordinarily concentrated in some ways, and fragmented and opaque in others. Rare Earth Exchanges® community members are well aware that China remained the top refiner for gallium, graphite, manganese, and rare earths with more than 90% of global supply in each case. Outside the dominant supplier, planned rare-earth refining capacity by 2035 equals only about two-thirds of diversified mined supply while planned magnet capacity equals only one-third. That imbalance helps explain why traders, brokers, and trading houses spent Q1 and Q2 2026 locking up offtakes, financing new corridors, and inserting themselves into government-backed reserves.

China’s controls kept the pressure on, ongoing. Rare Earth Exchanges has noted this year that export controls had already pushed Europe’s gallium and heavy rare earth prices to roughly five times Chinese domestic prices in some cases; by mid-2026 China’s flow of key heavy rare earths and gallium to Japan remained severely constrained. Chinese customs data reveal shipments of the seven covered rare earths to Japan fell 88% year over year in March and 82% in April. In other words, traders were not reacting to a theoretical problem. They were reacting to a live market dislocation.

A Busy Time

Traxys (opens in a new tab), based in Luxembourg, posted perhaps the busiest and most strategically significant first half of 2026 of any trader in the rare earth and critical minerals sector. Industry sources suggest the company's rare earth trading desk generated particularly strong results as geopolitical demand for secure supply accelerated. The firm emerged as one of the inaugural suppliers to the U.S. government Export-Import Bank of the United States (EXIM) Project Vault (opens in a new tab) strategic reserve, underscoring its growing role as a trusted intermediary for Western supply chains.

Commercially, Traxys moved aggressively across multiple critical mineral markets. In May, it signed a binding long-term supply term sheet with Arafura Rare Earths covering neodymium-praseodymium (NdPr) oxide and dysprosium-terbium (DyTb) oxide from the Nolans project in Australia. One month earlier, it executed a non-binding offtake framework with NioCorp for ferroniobium, scandium, and other products from the Elk Creek project in Nebraska. In February, the company expanded into battery materials through a graphite marketing memorandum of understanding with Sovereign Metals, followed in January by a landmark 10-year take-or-pay agreement securing 100% of Phase 1 lithium carbonate production from Lilac Solutions' Great Salt Lake project.

Taken together, the deals reveal a deliberate strategy extending well beyond rare earths. Traxys is positioning itself as a central merchant across the broader critical minerals ecosystem—using long-term offtakes, strategic marketing rights, and government partnerships to secure material flows before many new Western projects enter production.

Then, just after the period under review, it upsized its flagship revolving credit facility to $2.0 billion with support from a 30-bank syndicate, underscoring that banks now view critical-minerals merchandising as a scale business, not a side bet.

Based in Geneva, Switzerland, Mercuria’s (opens in a new tab) playbook looked different but no less strategic. It also joined Project Vault at launch, then moved deeper into physical critical-metals flows in Africa: in February it announced (opens in a new tab) its first copper-and-cobalt transaction with Entreprise Générale du Cobalt (opens in a new tab) after the DRC’s quota regime reshaped cobalt exports. TechMet (opens in a new tab), one of Mercuria’s important strategic linkages, reopened fundraising for up to $200 million, with Mercuria remaining one of its backers. The message was clear: Mercuria is not just trying to trade price spreads; it is positioning across finance, sourcing, and politically sensitive flows.

Hartree Partners (opens in a new tab), based in New York City, emerged as the third leg of the initial Project Vault stool alongside Traxys and Mercuria, with EXIM naming it as a servicing supplier. Hartree also put follow-on money (opens in a new tab) into Blue Moon Metals in April, backing a developer whose portfolio includes tungsten plus the U.S.-based Apex germanium-gallium-copper project. In a market where gallium and germanium were suddenly back on every defense and semiconductor watchlist, that was not an incidental placement.

Glencore plc (opens in a new tab), the Anglo-Swiss multinational commodity trading and mining company with headquarters in Baar, Switzerland, has sought to feed the stockpile-and-defense machine where it mattered. In February Glencore agreed to buy nearly $115 million of cobalt (opens in a new tab) from veteran Israeli trader Rami Weisfisch (opens in a new tab) for shipment to the United States in support of Project Vault. And there’s Glencore’s talks to sell a 40% stake in its Congo copper-cobalt mines (opens in a new tab) to the U.S.-backed Orion Critical Mineral Consortium (Orion CMC), showing how the big trading-miner hybrids were repositioning around the new American strategic-minerals complex.

A note: the Orion CMC is a $1.8 billion public-private investment platform launched by Orion Resource Partners (opens in a new tab), the U.S. International Development Finance Corporation (DFC), and Abu Dhabi's ADQ (opens in a new tab) to strengthen non-China supply chains for critical minerals. Targeting an eventual $5 billion fund, the consortium focuses on financing near-term mining projects and expansions—rather than early-stage exploration—across emerging markets in Southeast Asia, Central Asia, Africa, and South America to accelerate production of strategic materials such as rare earths, copper, cobalt, and lithium.

Especially considering the mounting tension with China, Japan’s trading houses were perhaps even more aggressive. Tokyo-based Sojitz (opens in a new tab), through Japan Australia Rare Earths, expanded its Lynas relationship twice in quick succession (opens in a new tab): samarium imports would begin in April 2026 and the product slate would broaden toward six medium and heavy rare earths by mid-2027. Then Sojitz said in March that it had signed an MoU with Lynas (opens in a new tab) to begin discussions on new mine exploration and development while also widening supply to Japan.

Multiple media outlets from Bloomberg to The Japan Times report the revised Japan-Lynas structure extended supply commitments to 2038 and embedded a $110/kg floor price for NdPr, with a substantial share of heavy rare earth output reserved for Japanese industry.

That is not everyday commodity trading. It is state-backed corridor building. Sumitomo (opens in a new tab), another Japan-based trading house, moved on both ends of the chain. MP Materials disclosed (opens in a new tab) via Securities and Exchange Commission that subsidiary Sumitomo Corporation of Americas (opens in a new tab) remains the exclusive distributor of MP’s NdPr oxide and NdPr metal to Japanese customers through the end of 2030.

Then in June 2026 Sumitomo announced (opens in a new tab) a strategic collaboration with Ucore Rare Metals to source feedstock for Ucore’s planned Louisiana Strategic Metals Complex and develop downstream offtake for separated products, especially middle and heavy rare earths for Japan while preserving material for North American and allied markets. That is classic sogo shosha (opens in a new tab) behavior: lock feedstock, lock distribution, and lock optionality.

South Korea’s POSCO International (opens in a new tab) behaved much the same way. In March it launched a 25 billion won corporate-venture-capital fund (opens in a new tab), put 8 billion won into a domestic separation-and-refining specialist, and laid out plans to secure roughly 4,500 tonnes of rare-earth product annually from Southeast Asian sourcing hubs before scaling toward 10,000 tonnes. In May it formalized a U.S. joint venture with ReElement to build a rare-earth separation, purification, and magnet-production complex in America, with phased capacity rising from 3,000 to 6,000 tonnes and total planned investment around $200 million. This was one of the clearest Q2 examples of a trading-oriented industrial group moving from procurement into full-blown value-chain construction.

Hanwa Co Ltd. (opens in a new tab) and Marubeni Corporation (opens in a new tab) also showed how the Japanese trading model is spreading beyond rare earths into adjacent chokepoints. Hanwa and JOGMEC signed an LOI (opens in a new tab) in February for up to $30 million of project-level equity in NextSource’s Abu Dhabi battery-anode facility, with potential Japanese offtake and logistics roles attached. And Marubeni teamed with Idemitsu, NSC, and Graphinex to develop a Japan-Australia natural-graphite-anode chain (opens in a new tab). These are not rare-earth deals narrowly defined, but they are absolutely part of the same strategic-metals trading rewrite, driven by the logic of the Great Powers Era 2.0.

Regional Signals

Chinese customs data indicate that the country's rare earth market tightened during the first half of 2026, with exports falling 6.4% year over year to 30,482.8 tonnes and imports declining 6.1% to 53,886.6 tonnes. While the figures alone do not explain overall supply-demand dynamics, they suggest a market increasingly shaped by policy and strategic management rather than normal price-driven trade.

Japanese and wider Asian reporting revealed how quickly buyers were adapting. In June Japan’s Ministry of Economy, Trade and Industry hosted (opens in a new tab) the Conference on Critical Materials and Minerals with the United States, Germany, France, Canada, Australia, the United Kingdom, Italy, and the European Commission, explicitly discussing permanent magnets as well as raw materials. Meanwhile, Japanese buyers hardened long-term offtake structures with Lynas, while broadening U.S.-Japan project cooperation across rare earths, lithium, copper, recycling, and gallium. In practice, the Japanese model is no longer just “invest upstream”; it is now “pay for resilience, even if that means floors, state agencies, and downstream alignment.”

In Korea the Seoul Economic Daily described POSCO International’s CVC-fund strategy not as a passive financial investment but as a deliberate move to secure heavy rare earths and then bolt that feedstock to U.S. processing and magnet capacity. South Korea, while participating in the U.S.-led FORGE bloc (opens in a new tab), simultaneously sought a hotline and joint committee with China to stabilize critical-mineral imports. That balancing act matters. Korea is trying to diversify without detonating the existing trade relationship that still feeds large parts of its industrial base.

Germany’s signal was political as much as commercial. The Financial Times reported (opens in a new tab) that Berlin was studying a Japanese-style trading house for critical raw materials, with names such as BMW and Rheinmetall involved in discussions and possible partial state funding on the table. At the company level, the thyssenkrupp (opens in a new tab) materials-trading arm continued advertising rare earth concentrate, lithium, and black mass supplied under long-term purchase agreements, even as the broader division’s spin-off into tk accelis (opens in a new tab) proceeded in June. In other words, Germany is moving toward a more organized trading-and-procurement posture, but it is still in institution-building mode rather than deal-flood mode.

France’s signal was industrial clustering. French reporting said the Lacq rare-earth “magnet valley” was beginning to pre-sell meaningful output, with more than 70% of future production already reserved, including contracts tied to Japan for dysprosium and terbium and supply to Stellantis. That dovetails with the U.S.-France deal announced this month under which USA Rare Earth took a minority stake in Carester to secure access to oxide output from Lacq. The Q1-Q2 takeaway was that France had moved from policy aspiration to real offtake positioning.

Europe also started building the market plumbing. The European Commission launched (opens in a new tab) its Raw Materials Mechanism in April to aggregate demand and connect buyers, suppliers, financiers, and storage providers, while EIT RawMaterials (opens in a new tab) and Metalshub (opens in a new tab) were working on a European critical-minerals pricing and trading platform. That may sound wonky, but it matters enormously. Traders can live with volatility. What they cannot scale easily around is the absence of trusted benchmarks and liquid counterparties. And as Rare Earth Exchanges has continuously reported, the rare earth and critical mineral sector continues to be opaque, with bespoke contracts.

Governments Procurement into Industrial Policy

Project Vault was the flagship American experiment of H1 2026. EXIM approved up to a $10 billion direct loan in February for the U.S. Strategic Critical Minerals Reserve, with total capitalization of roughly $12 billion including private investment, and said the structure would store essential raw materials in facilities across the United States through an independently governed public-private partnership. Initial OEM participants included Clarios, GE Vernova, Western Digital, Boeing, and others; initial named servicing suppliers were Hartree, Mercuria, and Traxys. The model aims to provide roughly 60 days of emergency mineral supply for U.S. manufacturers.

The July sequel made Vault even more important. The White House executive order of July 20 explicitly protected Project Vault and other U.S.-funded sources from being unintentionally impaired by tighter defense-sourcing rules, while Bloomberg Law reported in April that EXIM intended the reserve to be open to all trading companies, not just the trio named at launch. That makes Vault less a closed club than a new buying-and-buffering architecture around which more traders may gather.

The U.S.-Japan partnership was the other major Q1-Q2 state move. USTR announced (opens in a new tab) a U.S.-Japan Action Plan on Critical Minerals in March, explicitly pointing toward a future plurilateral trade arrangement supported by price floors or other measures. The White House described a Critical Minerals Action Plan aimed at increasing production and diversity, while Japan’s METI’s fact sheet tied that strategy to specific projects, including ReElement rare-earth recycling and Alcoa’s gallium recovery effort in Western Australia. This year’s final investment decision on the Alcoa gallium plant shows that some of those ideas are now starting to leave the whiteboard.

Australia built a parallel model with its Critical Minerals Strategic Reserve (opens in a new tab). Canberra said in January that the reserve would initially focus on antimony, gallium, and rare earth elements; the program combines offtake rights, selective stockpiling, and contracts-for-difference-like tools through an expanded financing facility. In May the government tied Arafura’s final investment decision directly to reserve support, with a non-binding commitment to secure 500 tonnes of rare earths from Nolans. Australia is not just subsidizing mines; it is trying to become a rules-based warehouse and supplier of last resort for allies.

The broader allied layer kept thickening. Treasury convened a critical-minerals finance ministerial in January with Japan, South Korea, Germany, France, Canada, Australia, India, Mexico, the United Kingdom, Italy, and the European Commission. Also, the United States pushed FORGE, the coalition built to accelerate secure and diversified critical-mineral supply chains, taking the FORGE chair in July after South Korea’s term. The state is not replacing traders in this market. It is recruiting them.

The Quiet Names and the Visibility Problem

REEx has introduced various articles tracking the traders and a broker map, involving firms such as HEFA Rare Earth Canada (opens in a new tab), G.E. Chaplin (opens in a new tab), Stanford Materials (opens in a new tab), plus thyssenkrupp alongside the bigger global houses. The Q1-Q2 2026 public record shows a sharp split between firms that issue press releases about projects and firms that simply keep selling product. On the latter side, G.E. Chaplin’s public materials still show a broad catalog of light and heavy rare earth oxides, metals, alloys, fluorides, concentrates, and powders; the company also lists dedicated rare-earth commercial staff and U.S. plus European operating points. HEFA Rare Earth Canada still markets direct supply from Baotou HEFA, including gallium and heavy rare earth products, while Stanford Materials continues to market rare earths and other specialty materials such as gallium and germanium to industrial and research users.

That distinction matters because, as we have reported, this market is still unusually opaque. The firms dominating headlines in 2026 are the ones announcing offtakes, JVs, reserves, and bank lines. But a large share of real business still happens through catalog supply, spot transactions, relationship-driven logistics, and small-volume qualification sales that never become public filings. There are perhaps a couple dozen specialized brokers in this space, some with an emphasis in Asia. Some firms even help secure product via gray markets.

Final Thoughts

The big H1 2026 lesson is that traders are no longer waiting for new mines to arrive and then figuring out how to move material. They are doing the opposite. They are pre-wiring supply chains before production exists by stitching together long-dated offtakes, reserve access, state-backed finance, and downstream customers. That is why Traxys could be in rare earths, graphite, niobium, scandium, and lithium at once; why Sojitz and Sumitomo are both expanding in Japan-facing rare-earth channels; why POSCO International is moving from sourcing to magnets; and why Mercuria, Hartree, and Glencore are showing up where stockpiles, cobalt, gallium, and germanium intersect. Specialized firms such as Chicago-based Pronto Recovery specialize in accessing hard-to-find gallium and germanium.

But the hard limit is still the middle of the chain. The Rare Earth Exchanges 2026 outlook makes that plain: the West is better today at announcing mines, reserves, and framework agreements than it is at building independent separation, metallization, alloying, and magnet ecosystems at scale. That does not make the current trader wave unimportant. It makes it transitional. Traders are buying time, creating optionality, and underwriting friend-shored corridors. They are not, on their own, solving the magnet gap.

The final caution is about deal quality. Many of the H1 2026 moves were real and consequential, but several were still term sheets, letters of intent, cooperation frameworks, or reserve structures with incomplete public detail. NioCorp’s arrangement with Traxys was expressly non-binding; Hanwa and JOGMEC’s UAE investment was an LOI; Project Vault’s storage, governance, and preferred-flow details were only partly public even months after launch. So the right read is not that a non-Chinese system is finished. The right read is that traders spent Q1 and Q2 2026 helping governments and industry lay down the rails for one.