Situation Briefing. Updated September 13, 2026.
The short version
China mines roughly 70 percent of the world's rare earths, refines about 90 percent, and makes 93 to 94 percent of the sintered permanent magnets. The bottleneck is processing and magnet-making, not ore.
The April 2025 licensing regime on seven medium and heavy elements is still fully in force. It was never suspended.
The tougher October 2025 measures are paused until November 10, 2026, which is 58 days away. They do not expire on that date. They resume unless Beijing extends the pause.
The most consequential rule is a 0.1 percent de minimis threshold that follows Chinese-origin material into goods manufactured entirely outside China. That is extraterritorial jurisdiction over molecules, and Beijing learned the technique from Washington.
Two clocks are about to collide: the pause lapses November 10, 2026, and a US defense ban on Chinese-origin rare earths takes effect January 1, 2027, 52 days later.
What this situation is
Rare earths are 17 metallic elements that are not actually rare in the ground. What is rare is the ability to separate them from each other, which is chemically difficult, environmentally punishing and industrially unglamorous, and the ability to turn the separated oxides into high-performance permanent magnets.
Those magnets are the point. Neodymium-iron-boron magnets are in electric vehicle motors, wind turbines, industrial robots, data centre cooling, hard drives, guided munitions, fighter aircraft and submarines. They have no drop-in substitute at comparable performance. In 2024, China exported 58,000 tonnes of rare earth magnets, enough for millions of vehicles or thousands of strategic military systems.
The concentration is severe at every stage and worsens as you move downstream. Global mine production reached roughly 390,000 tonnes of rare earth oxide equivalent in 2025, with China accounting for 270,000 tonnes, about 69 percent. The IEA estimates China performs around 91 percent of global refining and separation and about 94 percent of sintered permanent magnet production.
China's position strengthens at each downstream stage. Sources: IEA, GlobalData, industry estimates.
Two elements matter more than the rest. Dysprosium and terbium are what keep a magnet stable at high temperature, which is exactly what a motor or a missile requires. They are heavy rare earths, the hardest category to source outside China, and they are the ones Beijing has controlled most tightly.
How the controls were built
This was not a single announcement. It was a three-year architecture, and the sequencing is deliberate.
August 2023: gallium and germanium placed under export controls, a first test of the instrument on semiconductor inputs.
April 2025, Announcement No. 18: mandatory case-by-case MOFCOM licensing for seven medium and heavy elements, samarium, gadolinium, terbium, dysprosium, lutetium, scandium and yttrium, plus all metals, oxides, alloys, compounds and the permanent magnets containing them. The effect was immediate. European dysprosium reached up to six times the Chinese domestic price and several automakers cut production within weeks.
October 2025, Announcements No. 61 and 62: the escalation. No. 61 extended control extraterritorially through a 0.1 percent by-value de minimis rule, meaning a magnet made entirely outside China falls under Chinese licensing if Chinese-origin rare earth content reaches that threshold. No. 62 controlled the technology and know-how itself, the processes for extraction, separation and magnet manufacture.
November 2025: after the Xi-Trump meeting, MOFCOM Announcement No. 70 suspended the October package until November 10, 2026. Washington suspended its Affiliates Rule for the same period. A mutual stand-down, not a settlement.
2026: the architecture kept expanding during the truce. The January licensing catalogue added rare earth compounds and silver. On March 31, State Council Order No. 834 promulgated China's first dedicated supply chain security framework, integrating export controls, countermeasures and data security into one regime. Early in the year, new dual-use restrictions were aimed at Japan, one of the few countries outside China that makes rare earth permanent magnets.
Read that sequence again and the pattern is clear. The suspension covered the headline measures while the underlying legal machinery was extended, codified and institutionalised. Nothing was dismantled. The licensing infrastructure was switched off, not removed.
Why the 0.1 percent rule is the real instrument
The de minimis threshold deserves attention that tonnage figures do not.
A conventional export control says: you may not ship this item from our territory without permission. A de minimis rule says: any item anywhere in the world containing more than a trace of our material remains subject to our permission, forever, through every subsequent transformation.
Applied at 0.1 percent by value, that captures nearly everything. A server's cooling fans, the sputtering targets used in chip fabrication, the magnets in hard drives. The permit requirement attaches at component transfer, not at final import of the finished system.
This is precisely the logic the United States has used for decades in semiconductor export control, where American-origin technology content brings foreign-made chips under US jurisdiction. Beijing has copied the architecture and pointed it at the one supply chain where it holds the commanding position.
And there is a refinement buried in the mechanism. The threshold is value-based, which means Beijing can widen or narrow the effective scope of its own rule by moving domestic Chinese prices. The control has a dial on it.
Export controls used to be about borders. The de minimis rule makes them about provenance, and provenance follows an atom forever. A supply chain cannot diversify away from a jurisdiction that claims authority over material it once touched.
Why diversification has not worked
Every government exposed to this has announced a response. The US-Australia Critical Minerals Framework, Inflation Reduction Act and CHIPS Act provisions, Defense Department funding for heavy rare earth magnet projects, the EU's Critical Raw Materials Act targeting 40 percent domestic processing by 2030, and expanded Five Eyes coordination.
Six months into the suspension, analysts concluded the available data did not suggest readiness, with diversification described as gradual and capital-intensive and China expected to hold its position in processing and magnets through the decade.
Three reasons the money has not translated.
First, the binding constraint is process knowledge, not deposits. China's advantage rests on separation expertise, environmental tolerance for refining, scale economies and an integrated magnet ecosystem. Announcement No. 62 targeted exactly this by controlling the technology transfer, which is an attempt to freeze competitors at their current capability.
Second, the historical precedent is discouraging. The 2010 rare earth shock triggered a surge in related patenting and some downstream production outside China, and the gap still did not close quickly. That episode is the closest thing to a controlled experiment available, and it ran for fifteen years.
Third, capacity is thin where it matters most. Noveon Magnetics was the only notable US sintered NdFeB producer as of early 2026, and meaningful new heavy rare earth capacity was not scheduled to enter the chain until 2027.
The price data confirms it. Dysprosium oxide in Europe traded at roughly 4.9 times the Chinese domestic price in late August 2026. That spread is the measured cost of supply sovereignty, and it has not compressed.
Who is exposed, and unevenly
Licensing has not been applied uniformly, which is itself a form of leverage.
When China eased restrictions in November 2025, magnet exports rose 13 percent overall. Shipments to Europe jumped 60 percent year over year while US imports fell 11 percent and have never recovered to pre-restriction levels. The yttrium figures are starker: 17 tonnes to the United States across eight months from April 2025, against 333 tonnes in the eight months prior.
That divergence is a policy choice. Differential licensing lets Beijing apply pressure to one government while giving others a reason not to join a coordinated response, the same logic the Houthis applied to Saudi shipping at Bab el-Mandeb. Exempt most of the market and you never face a unified opponent.
The United States is the primary target and the least supplied. Europe has fared better on volume and still pays the heavy rare earth premium. Japan holds rare non-Chinese magnet capability and was specifically targeted in early 2026, which is not a coincidence. Automakers and wind developers absorb cost. Defense primes face something worse than cost, which is a compliance requirement they may not be able to satisfy.
Where it stands now
The April 2025 regime is live. Foreign defense buyers are still being denied licences for the seven controlled categories. The October 2025 package is paused until November 10, 2026, with the technology controls and the extraterritorial rule both waiting behind that date. The separate suspension covering US-focused dual-use licensing of gallium, germanium, antimony and graphite runs to late November 2026.
If the pause lapses, the controlled list expands from seven elements to twelve, adding holmium, erbium, thulium, europium and ytterbium, along with their alloys, oxides and compounds.
What to watch
November 10, 2026, and the framing around it. Treat it as a planning horizon, not a resolution date. Three outcomes: a further extension, which signals Beijing values the trade relationship more than the leverage; a clean snapback; or a redesigned framework that preserves leverage while reducing friction, which is the most likely and the hardest to model.
The 52-day collision. This is the structural risk almost nobody is pricing. US Defense Federal Acquisition Regulation restrictions expand on January 1, 2027 to cover the entire NdFeB supply chain from mining through finished magnets, forcing defense buyers toward traceable non-Chinese sources. That lands 52 days after the pause lapses. In the worst case, Washington mandates non-Chinese sourcing at almost exactly the moment Beijing reasserts extraterritorial claims over Chinese-origin content embedded in third-country production. Defense primes would face a legal requirement with no physically compliant supply route.
Dysprosium and terbium spreads, not aggregate tonnage. Total rare earth volumes tell you almost nothing, because the constraint is a specific element at a specific purity in a specific form. The European-to-Chinese price ratio on dysprosium oxide is the honest gauge. If it compresses, diversification is working. It has not compressed.
Whether the technology controls in Announcement No. 62 return. Materials controls raise costs. Technology controls determine whether a competing industry can be built at all. The second is strategically decisive and gets a fraction of the attention.
Licensing asymmetry between the US, Europe and Japan. Widening divergence means the coalition is being split successfully. Convergence means it is holding.
The deeper point is that this was never a commodity story. Ore is abundant and the deposits are widely distributed. What Beijing controls is the midstream, the know-how and, since October 2025, the legal claim over its own atoms wherever they end up. Countries have spent two years funding mines to solve a problem that does not live in the ground.
You can open a mine in five years. You cannot open a jurisdiction.
This Situation Briefing is updated monthly. Last revised September 13, 2026.