MACEDON STRATEGIC INTELLIGENCE
THE MACEDON ASSESSMENT
The gap between what is announced and what can be executed.
China's rare-earth dominance rests on expertise built over decades, not merely mines or refineries. Beijing is tightening enforcement around the channels through which that knowledge moves.
The West is answering China's rare-earth dominance with a build-out: finance the mines, the refineries, and the magnet plants, and the processing bottleneck is supposed to disappear. The harder constraint is human. A refinery runs on process chemists, hydrometallurgists, and the tacit skill of solvent extraction, the complex separation step China has mastered and Western firms have repeatedly not. China performs roughly 90 percent of the world's rare-earth processing, and by the IEA's count 91 percent of refined output and 94 percent of sintered permanent magnet production in 2024. In June 2026, MOFCOM Announcement No. 26 did not invent a new control on knowledge; it built an enforcement architecture around an existing one, aimed at the person-to-person channels that are the slowest layer for the West to rebuild. Capital moves in years. Industrial memory moves in decades. That asymmetry is the real chokepoint.
01The announcement
The Western answer to Chinese rare-earth dominance is a build-out. Governments and companies are financing domestic mines, refineries, and magnet plants, and analyses such as the CSIS processing-hubs framework score the United States and its allies as capable of standing up alternative capacity. The implicit premise is that with enough capital and political will, the processing bottleneck can be built out of existence.
The scale of what has to be replaced is not in dispute. CSIS puts China at about 90 percent of global rare-earth processing while it mines only about 60 percent of the ore, the mark of a midstream built on purpose. The IEA, scoping to the magnet rare earths that matter most, puts China at 91 percent of refined output and 94 percent of sintered permanent magnets in 2024, up from around half in 2005. For the heavy rare earths specifically, the separation step happens almost entirely in China; the USGS notes that commercial-scale heavy-rare-earth processing outside China is only now being developed, with at least five US companies working on it as of 2025.
02The gap
Capital builds plants. It does not build chemists. Three layers the build-out framing omits.
- The binding expertise is specific and scarce. The hardest step in rare-earth separation is solvent extraction, a complex, tacit, hard-won skill. CSIS states plainly that China holds unmatched expertise in it, while Western companies have struggled on workforce capability, research and development, and environmental regulation. A refinery without that expertise is a building, not a capability.
- Expertise runs on a decade clock, not a capital clock. Facilities can be financed and constructed in a few years. The process engineers and hydrometallurgists who run them at yield, and the institutional learning-by-doing they carry, take decades to cultivate, and the teachers to rebuild them in the West are thin. China sustains that pipeline directly: Reuters counts at least eleven Chinese universities with dedicated rare-earth programs enrolling more than 500 students a year, and more than forty specialist laboratories. The United States awarded about 200 generalist mining and metallurgical engineering degrees in 2023, and no specialized rare-earths undergraduate degree exists outside China. One processor put it plainly to Reuters: a Chinese graduate is immediately productive, while a new hire elsewhere needs about three years of training.
- The enforcement now targets the knowledge channels. In June 2026, MOFCOM Announcement No. 26 established a whole-society reporting-and-reward mechanism to enforce China's existing export controls on strategic-mineral technology, issued June 24 and effective July 1, 2026. Its thirteen enumerated categories of reportable conduct expressly cover the illegal transfer of controlled strategic-mineral technology through knowledge transfer, joint R&D, employment, consulting, IP licensing, investment, testing, and assistance. The technology control was already on the books; Announcement No. 26 builds the enforcement architecture around it, directing public reporting and rewards at the person-to-person channels through which industrial knowledge actually moves. The conclusion is not that China newly controls expertise, but that it is making its existing controls more detectable and enforceable across exactly those channels.
03The window
A refinery financed today can be commissioned in a few years. The expertise to run it at design yield cannot be commissioned on the same clock, and the pool it would be hired or learned from sits overwhelmingly inside China, whose new enforcement reaches the very channels that pool would move through. The build-out addresses the layer money can buy and leaves the layer money cannot.
The lag is already visible in the numbers. CSIS, reviewing the year since China's April 2025 export restrictions, draws the distinction between distance and displacement: announced and financed capacity is one thing, actual rare-earth and magnet output is another, and displacement remains modest. New magnet manufacturing capacity coming online in the summer of 2026 begins to reduce reliance on China, but self-sufficiency, in CSIS's words, remains a long road, because translating announcements into production takes years.
04The exposure map
The exposure concentrates wherever a processing or magnet project is underwritten on a capital-and-timeline basis that assumes the expertise will be there when the plant is. It runs through the Western refinery and magnet build-out, the allied hub strategy, and every downstream manufacturer, in defense, electric vehicles, and electronics, counting on that capacity arriving on schedule and at yield. The binding variable is trained human capital multiplied by time, not plant count. This is the layer beneath the material and the resource: No. 01 traced the material chokepoint in rare earths, No. 04 the resource-versus-supply gap in lithium, and No. 05 goes one step deeper, because the midstream runs on expertise that took decades to accumulate and is now the object of active Chinese enforcement. Mapping which projects have a credible expertise and training pipeline, and which have only financed steel, is the analysis that pays.
05The signals to watch
- Whether Western processing and magnet projects reach design yield, not merely mechanical completion, and how long the ramp takes.
- Reports and enforcement activity under MOFCOM Announcement No. 26, especially cases involving personnel, consulting, or joint R&D.
- The formation of Western training pipelines: university programs, apprenticeships, and retained expertise in solvent extraction and hydrometallurgy.
- The movement of experienced personnel, and any restrictions or incentives shaping where that expertise can go.
- The gap between announced hub capacity, on CSIS-style scoring, and demonstrated at-yield output.
06Sources
- CSIS, "Developing Rare Earth Processing Hubs: An Analytical Approach," Gracelin Baskaran and Meredith Schwartz, July 2025.
- CSIS, "Rare Earth Export Restrictions One Year Later," Gracelin Baskaran, April 2026.
- IEA, Rare Earth Elements, executive summary, 2024 data (91 percent refined output, 94 percent sintered permanent magnets).
- U.S. Geological Survey, Mineral Commodity Summaries 2026, rare earths (heavy).
- MOFCOM Announcement No. 26 of 2026, PRC Ministry of Commerce, issued June 24 and effective July 1, 2026.
- Reuters, "A bachelor's in rare earths? In China, there are schools for that," June 2026.
- The Macedon Assessment No. 01 (the material layer) and No. 04 (the resource layer).