Rare Earth Supply Chain
Geopolitical Risk Analysis

The mechanism is called the Foreign Direct Product Rule. Washington has used it to prevent chips and chipmaking equipment containing American technology from reaching Chinese buyers, even when the products are manufactured in third countries like the Netherlands or Japan or South Korea. The core claim is extraterritorial: if our technology is inside your product, we get a say in where that product goes.

Announcement No. 61 took that claim and planted it in rare earth soil. Under the new rules, a foreign-manufactured product that contains Chinese-origin rare earth materials above a 0.1% value threshold requires an export license from Beijing. Products made anywhere in the world using Chinese rare earth processing technology require the same license. Chinese-origin rare earths transiting through third countries require the same license. Georgetown University's CSET translated the full text; the operative phrase is "manufactured in foreign countries, and that contain, integrate, or have mixed in with them any rare earths originally produced in China." That is not aspirational language. It is a regulation with a number attached.

Consider what this means for a transaction that has nothing to do with China on its face. A Japanese magnet manufacturer buys dysprosium that originated in China, presses it into a permanent magnet, and sells that magnet to a German wind turbine company. Under Announcement No. 61, the Japanese company needs Beijing's approval to make that sale to Germany. Mayer Brown's legal team described the scope as "materially expanding compliance touchpoints for non-PRC manufacturers and distributors." ASML started internal assessments almost immediately.

The symmetry with the American semiconductor FDPR is obvious and presumably intentional. Washington claims authority over foreign-made chips with American technology inside them. Beijing now claims authority over foreign-made magnets with Chinese materials inside them.

Announcement No. 61 was partially suspended after the Busan talks between Trump and Xi in late October 2025. The FDPR provisions, five newly added elements (holmium, erbium, thulium, europium, ytterbium), and some dual-use licensing requirements were put on ice for one year through November 2026. Media reported this as a rare earth détente. What got much less attention is that Announcement No. 18, which went into effect on April 4, 2025, was not part of the Busan deal at all. That announcement placed seven medium and heavy rare earth elements (samarium, gadolinium, terbium, dysprosium, lutetium, scandium, yttrium) and all their derivative products under case-by-case export licensing. It ran continuously through 2025 and is still running. Trump said the rare earth issue was settled. It was not settled. FDD called Busan a "tactical pause," which is closer to what happened. Washington paid for the pause by shelving the Bureau of Industry and Security's Affiliates Rule, a chip-domain tool that would have extended export licensing to affiliates of sanctioned Chinese entities. A concession in semiconductors bought a temporary step-back in rare earths, while the April controls continued operating underneath.

The Ford situation from mid-2025 is where all of this stops being abstract. Jim Farley said at the Aspen Ideas Festival in June that Ford had shut down plants for three weeks because they could not get high-powered magnets. He was specific about what the magnets go into: audio speakers, seat motors, wiper motors, door motors. "Day to day" and "hand-to-mouth" were his descriptions of the supply picture. The Chicago Assembly plant, building the Explorer, went down for a week in late May. Farley said more than one plant was affected but did not name them.

The parts he listed are not electric vehicle components. They are in every internal combustion vehicle in production. Wiper motors. Seat adjusters. Door actuators. Fuel injectors. The rare earth supply chain problem that gets framed as a future electrification risk is already a present-tense conventional automobile risk, and this point keeps getting buried under the EV and wind power narrative.

China did not embargo Ford. The April 4 licensing regime requires case-by-case approval for each export shipment. Some American companies were told the process takes 45 days. There is no statutory guarantee attached to that number. A modern automotive production line running lean carries maybe a week of safety stock on small components. Forty-five days of wait time with an uncertain outcome is a shutdown by another name.

The Geneva talks in May produced a truce that supposedly included restoring rare earth supply channels. Farley said weeks later that nothing had improved. Applications in, being processed one at a time. The licensing system is fully WTO-compliant on its face because every sovereign state has the right to license dual-use exports. The system can differentiate between applicants. It can be tightened or loosened by adjusting processing speed without any change in regulation. And it gives Beijing complete deniability.

This has a specific implication for the diversification projects everyone is talking about. Even if new processing capacity gets built in the West, it will probably lose money during any period when Chinese supply is flowing normally. The capacity becomes insurance that sits on the balance sheet waiting for a crisis. Insurance costs money. The global rare earth market does under $4 billion a year in revenue. The question of whether a $4 billion market can support distributed insurance capacity across multiple continents has not gotten a serious answer from anyone. The U.S. government putting $1.6 billion of public equity into USA Rare Earths in January 2026 is an implicit admission that private capital alone will not build this.

Noveon Magnetics, as of the October 2025 CSIS report, was the only company manufacturing rare earth permanent magnets in the United States. One company. Noveon and Lynas signed an MOU the same week to cooperate on scaling domestic magnet production. China exported 58,000 tons of rare earth magnets in 2024. Industrial-scale separation facilities operating outside China: Lynas in Kuantan, Malaysia; Mountain Pass in California, primarily light rare earth concentrate with separation capacity under construction; NPM Silmet in Estonia. IEA puts the standard project timeline at about eight years from initiation to production. Many magnet-grade ores carry uranium and thorium. Countries with the regulatory infrastructure and physical facilities to handle radioactive processing waste are scarce. An optimistic assessment for a complete mine-to-magnet chain in the U.S. or EU is mid-2030s.

The financing structure underneath these projects has a problem that does not get discussed much. Processing capacity needs downstream offtake agreements to secure lending. Offtake buyers compare prices. Chinese rare earth supply is stable and cheap almost all the time. An automaker has no commercial reason to sign an expensive long-term contract with a non-Chinese supplier when Chinese supply is working fine. The demand for non-Chinese supply spikes only when Chinese supply breaks. Projects outside China are in the strange position of needing geopolitical instability to generate their demand. Their pitch to investors is implicitly a pitch about conflict scenarios. This is not a comfortable investment thesis, and the difficulty of raising private capital for these projects reflects that discomfort.

Northern Myanmar, Kachin State. Ion adsorption clays with high heavy rare earth content. One of the most important heavy rare earth sources outside China proper, though you would not know it from most public analysis, which tends to mention it in passing if at all. Ore crosses the border into Chinese separation facilities and enters the global supply chain from there. Mining operations are entangled with the Kachin Independence Organization and other armed groups. Logistics run through informal border channels. When fighting in northern Myanmar escalated in late 2023, ore shipments to China dropped, and dysprosium and terbium prices jumped. Global heavy rare earth supply is hanging on two things at once: Chinese willingness to process and armed stability in a region of Myanmar that has been in conflict for decades. A local commander's battlefield decision in Kachin State can spike permanent magnet spot prices faster than any policy document coming out of Washington or Brussels. This kind of micro-instability at the physical origins of supply chains does not fit neatly into the frameworks that most geopolitical analysis uses, which tend to focus on capitals and ministries rather than border crossings and jungle roads.

On the demand side, the picture is getting crowded. Defense, EVs, wind power, and AI infrastructure are all pulling on rare earth magnets at the same time, and none of these sectors is shrinking. Chris Berry, an independent analyst who covers battery metals, has said that with European defense rebuilding, AI data center capital spending still expanding, and robotics demand growing, he does not see rare earth prices coming down in 2026 or 2027. When supply is tight, defense gets priority because governments can invoke legal tools like the Defense Production Act. The commercial sectors eat whatever is left. Automakers pay more or wait longer. Wind farm schedules slip. Data center component deliveries stretch. The export controls are nominally aimed at defense supply chains, but the demand-side squeeze means that green energy and tech may absorb most of the actual cost, because defense has statutory priority and budget backing that commercial buyers lack.

Beijing's planners can see this transmission mechanism as well as anyone else can. Whether rare earth controls are designed to simultaneously pressure Western defense, slow electrification, and constrain AI buildout, or whether those are incidental side effects, is not something you can determine from outside. The structure of the effect is the same either way.

IEA's 2025 Critical Minerals Outlook reported that for 19 of 20 important strategic minerals, China is the largest refiner, with an average market share of 70%, and that this concentration has increased in recent years. In sintered permanent magnets specifically, China's share went from about 50% two decades ago to 94% now. Diversification efforts are real. They are also being outrun. China is not waiting; it is expanding and upgrading an industrial chain that already has three decades of scale and learning curve behind it. ODI projects China still supplying about 80% of battery-grade rare earths in 2035. The target that diversification is chasing is moving away faster than the chase is closing.

On the environmental side: each ton of rare earth mined produces up to 2,000 tons of toxic waste, some radioactive. The tailings lake outside Baotou is probably the most photographed environmental disaster in the mining industry. Chatham House published a piece in March 2026 asking whether the rush to diversify rare earth supply would lead to lowered environmental standards, particularly in countries where environmental governance is already weak. Their assessment was that it probably would. In countries with functioning court systems, environmental lawsuits can stall a rare earth project for years. The processing system that China built came at an environmental cost that took decades to accumulate. Building equivalent capacity faster while maintaining higher environmental standards is an arithmetic problem that has not been solved.

Everything described above happened already. Ford's three weeks of downtime. China's FDPR. The sixfold European price gap. One American magnet manufacturer. The Busan deal that left Announcement No. 18 untouched. The enforcement bureau hiring while the policy was on pause.

China controls separation chemistry and magnet manufacturing, the two segments of the value chain where replication is slowest and alternatives are thinnest. Export tonnage going up does not change this. Building a parallel capability outside China takes about ten years on the optimistic end. Whether the political will and capital commitment required to sustain that effort over ten years will actually materialize is an open question. Rare earths are a small, technical, unglamorous market. They do not hold headlines the way oil does, or chips, or AI. The attention will drift. It always does with topics that are critically important but hard to photograph. Then the next disruption hits and the same questions get asked again about what happened in the meantime.