From Critical Minerals to Critical Components: AI, China, and the New Geography of Chokepoints

By J. M. Lindstrom

Geopolitical contests have long turned on control of strategic resources and the subtle power plays built around them. From mines in the Democratic Republic of the Congo to rare earth magnets destined for the Pacific seabed, mineral extraction has become an imperative step in the supply chain for a new metric of national power. No single nation monopolizes extraction: reserves sit in Australia, Vietnam, Brazil, Russia, and the United States alongside China1, and Washington's own stockpile accounting now tracks this dispersion directly2. What is concentrated is the next link in the chain: separation and refining, where China accounts for roughly 90 percent of global capacity3, and magnet manufacturing, where its share approaches 94 percent4. This dominance in processing, not in raw deposits, is what has driven the EU and the United States to chase refining capacity of their own. But a closer look up the supply chain reveals the deeper problem: once refined minerals are secured, leverage simply shifts to the next chokepoint, critical components.

Geopolitics increasingly demands quantifiable metrics. Militaries and economies once filled that role; as the world has grown more economically entangled, critical minerals and export-control cycles have taken their place. Microchips preceded rare earth minerals as a metric of chokepoint power, paired with rare earth magnets and processing technology; the underlying pattern is that the chokepoint rises up the supply chain over time. A decade of quantitative research in Resources Policy5 has tried to measure exactly how much power this concentration confers. Zhang and colleagues6 found Chinese market power over rare earths rose sharply through the 2000s and 2010s even as Beijing's own export policies produced mixed results for its stated goals. Barteková and Kemp's regional comparison7 found no other bloc not the EU, not Japan, not the United States had built a comparably durable national strategy for securing supply. Massari and Ruberti's earlier survey8 traced how "critical" status itself became a policy category only after China's early-2010s restrictions forced the issue. More recent work has tried to model the market's actual mechanics: a dynamic partial-equilibrium model of metal markets9 built specifically around rare earths, a catastrophe-progression early-warning framework10 for supply security, and a 2023 comparative index11 ranking China, Australia, Russia, and India on rare earth competitiveness, which still placed China well ahead on every measured dimension but raw reserve share.

Minerals have served as a pressure point on the world economy, one felt more acutely by technology-intensive Western economies than by most of the developing world, the Asian Tigers being a notable exception. Deng Xiaoping is widely quoted as telling aides touring Baotou in January 1992 that "the Middle East has oil, China has rare earths," a line China watchers have repeated ever since12 as evidence Beijing always intended to weaponize the resource. That reading is contested: the quote appears to have entered Western reporting only in 200913, and at least one scholar who traced contemporaneous Chinese sources argues the original remark was about something closer to the opposite point. Wübbeke's policy history14 of China's rare earth industry likewise complicates the simple weaponization story, attributing Beijing's restrictions mainly to domestic environmental and resource-conservation politics rather than grand strategy. Whatever Deng meant in 1992, Beijing's subsequent conduct has been unambiguous: China curtailed rare earth exports to Japan for roughly two months in 201015 amid a dispute over a fishing-boat captain near the Senkaku/Diaoyu Islands. Western powers hold substantial raw ore reserves but lack the processing capacity to convert them into usable material, a chokepoint in which much of the remainder outside China, including Indonesian nickel processing, is itself owned by Chinese firms operating abroad16, according to the IEA's regional accounting. That this concentration has since hardened into deliberate state policy is visible in Beijing's August 2025 shift to centralized, traceable state control17 over rare earth mining and smelting.

The current fight over rare earths is not the first. When China first tightened export quotas in 2010, the EU, the United States, and Japan took the case to the World Trade Organization; a WTO panel ruled against Beijing in 201418, finding the quotas violated China's accession commitments. Legal scholarship on the case19 argued the ruling exposed a real gap in WTO law: export restraints on raw materials were treated far more leniently than import barriers, even though both distort trade. That asymmetry is part of why Beijing's current approach targets technology and downstream products rather than raw ore directly, a structure the OECD's own trade-policy research on export restrictions20 and its later inventory of restrictive measures worldwide21 flagged as a structural loophole years before this cycle began. The WTO's own economists have separately modeled how such trade restrictions distort global welfare22 more broadly, a framework that applies just as well to today's foreign direct product rule as it did to 2010's quotas.

China's late-2023 and 2024 controls targeted not ore exports but processing technology, a dynamic mirrored in the competition between the United States, Taiwan, and China over microchips. CSIS's timeline of the escalation23 runs from October 2022 chip controls through the July 2023 gallium and germanium licensing regime; export volumes of both metals collapsed within months24 of that first round, according to USITC trade data. Antimony followed in 2024, and rare earth processing technology and the elements themselves followed in 2025, the latter accompanied by an explicit foreign direct product rule and a 0.1 percent de minimis threshold25 reaching into products manufactured outside China. One independent timeline of the full three-year sequence describes it as a controlled climb up the value chain26 rather than a single embargo. Analysts at CSIS have called the October 2025 measures the first Chinese application of foreign direct product rule logic to defense-relevant supply chains specifically27, while MERICS's running series on the controls28 tracks the same measures from the Chinese-industrial-policy side rather than the trade-law side. Those October 2025 measures, MOFCOM Notices 61 and 62, were set to take effect December 1, 2025, but Beijing suspended their implementation for one year on October 30, 202529, as part of a broader trade truce with Washington. FDD's assessment of the pause cautioned that suspension is not repeal30; Beijing retains the legal architecture and the option to reactivate it on short notice, a reading Resources for the Future's own analysis of the pause shares, arguing Beijing likely prefers reversible leverage31 to a permanent embargo. The U.S.-China Economic and Security Review Commission's own scenario modeling of a full chokepoint-stacking event32 treats exactly this kind of reversible-but-real leverage as the central risk facing the defense industrial base.

The newest and largest source of demand for rare-earth-dependent components is AI compute and data center infrastructure motors for cooling systems, advanced semiconductor nodes, and rare-earth-adjacent materials in high-bandwidth memory and chip packaging. FP Analytics has traced direct linkages between gallium, germanium, tungsten, and magnesium and specific AI hardware bottlenecks33, while CSET's foundational mapping of the chip value chain34 shows those bottlenecks are not new, only newly urgent. CSET's separate primer on why AI chips are a distinct, largely non-substitutable category35 helps explain why the component stage, not the mineral stage, increasingly carries the leverage; its 2023 policy brief on securing semiconductor supply chains36 proposed a protect-and-promote framework for exactly this kind of allied chokepoint defense. In practical terms, a data center's procurement list does not read "germanium," but "high-speed transceivers," a component made of it. China's Digital Silk Road strategy reinforces this pattern: Beijing has worked to close the supply chain on both ends, processing upstream and downstream AI infrastructure alike. A peer-reviewed chapter from the Hungarian Institute of International Affairs37 maps this data-center-and-cable strategy in detail, and a 2025 study in the Advanced Social Science Archive Journal38 extends the same architecture across South, Southeast, and Central Asia. Separate research has documented comparable patterns in Africa39 and, further afield, case studies in Brazil and South Africa40, where the DSR functions simultaneously as a development mechanism and a strategic-risk instrument. The Council on Foreign Relations' own case study of the AAE-1 submarine cable route through Southeast Asia41 shows the same pattern at the infrastructure level, though the extent to which any of this reflects coordinated strategy versus convergent opportunism remains genuinely contested among specialists. China's approach blends civilian and military applications of this infrastructure, and its 14th Five-Year Plan's dual circulation doctrine42 further insulates a secondary domestic flow of rare earth minerals and components from the effects of any export restriction Beijing might itself impose.

This leverage is not evenly distributed; it concentrates on specific regions. Myanmar has emerged as an unlikely but critical node: China's own supply of heavy rare earths increasingly comes from ionic clay mining in Kachin State43, territory the Myanmar government does not fully control. The extraction method itself is not new; the same ionic clay techniques were pioneered domestically in southern China44 before spreading across the border as Chinese regulators tightened rules at home. Myanmar accounted for roughly 57 percent45 of China's rare earth imports as of mid-2025, and the Kachin Independence Army's 2024 seizure of mining sites responsible for about half of world heavy rare earth production triggered price spikes that exposed how much of Beijing's own supply chain runs through territory it does not control. Satellite-imagery documentation of the extraction footprint from 2010 to 202546, compiled by USGS, shows the scale of that dependency in physical terms. Not everyone reads this as Chinese vulnerability, though: East Asia Forum has argued Myanmar's leverage is overstated47 relative to China's continued command of refining, since ore without a buyer willing to process it is worth little. The broader pattern nonetheless holds: extraction in states with weak governance capacity, refining concentrated in China, and downstream digital infrastructure built by Chinese firms in many of the same regions that supply the upstream inputs.

The argument that leverage now sits in components rather than raw minerals is not just theoretical. It is visible in the Tomahawk, JASSM, and Patriot interceptor stockpiles the United States has drawn down during its 2026 war with Iran. An independent CSIS analysis48 found several of these inventories will take three or more years to rebuild to pre-war levels, a bottleneck driven not by raw material scarcity but by the pace at which guidance systems, seekers, and precision-machined components can actually be manufactured. Senator Mark Kelly's May 2026 warning49 that some timelines could stretch into years drew public pushback from the Pentagon, underscoring how politically sensitive the component bottleneck has become. The episode is a live illustration of this piece's central claim: mineral scarcity is rarely the binding constraint anymore. Component manufacturing capacity is.

The West has scrambled to respond to China's processing dominance, primarily through the Inflation Reduction Act's critical mineral provisions, the CHIPS Act, and a small number of high-profile deals. The Department of Energy's own 2023 critical materials assessment50, finalized that August via Federal Register notice51, set the baseline list Washington still works from. The most significant deal since is the Department of Defense's July 2025 partnership with MP Materials, in which the Pentagon became the company's largest shareholder through a \$400 million equity stake52, a structure confirmed in the company's own SEC filing53 and backed by a \$150 million DoD loan54 for processing upgrades. Independent analysis from the Payne Institute55 at the Colorado School of Mines notes the deal marks the first time the Pentagon has taken an equity position of this kind, a departure the Bipartisan Policy Center has called a precedent for federal industrial policy56, not just for minerals. Washington's other lever is the Commerce Department's Entity List: dozens of Chinese firms57 have been added over the past two years, including a further round58 targeting semiconductor-adjacent suppliers and another expansion in September 202559 a reminder that Washington plays the weaponized-interdependence game too, just from the demand side rather than the supply side. The broader friend-shoring strategy has nonetheless been slowed by bureaucratic permitting delays and the absence of full value-chain buildout. The Government Accountability Office has repeatedly found DoD has not fully implemented statutory requirements60 Congress already passed, a finding that echoes GAO's earlier 2022 review61 of federal recovery and substitution efforts and its 2024 assessment62 of DoD's own reduction options, as well as a separate review of defense industrial base risk63. The European Union has pursued a parallel path through its Critical Raw Materials Act64, which sets 2030 benchmarks of 10 percent domestic extraction]{.underline} [and]{.underline} [40 percent processing65 targets that, per the Council of the EU's own tracking66, describe an ambition rather than a present capability. Jones Day's legal analysis of the Act's Strategic Projects mechanism found permitting timelines still lag China's pace67 by years, even after UNECE's 2024 assessment called it the most ambitious framework of its kind68 to date. A year on, CSIS's own retrospective69 found the combined U.S. and EU response had restored some resilience but had not closed the underlying speed gap with Beijing's state-directed capital deployment.

The leverage China has built through this supply chain is, at root, a higher-order expression of an infrastructure-for-resources exchange the country has used elsewhere to build alignment. The potential leverage of AI-era components amounts, in short, to a monopoly and a formidable geopolitical lever. Should China constrain gallium or germanium flows at the height of the AI compute race, it could trigger a compute chokepoint event in which other nations fall behind for lack of components, while China's dual circulation strategy allows it to continue functioning domestically. Recent theoretical work on how both Washington and Beijing construct minerals as security issues70 argues this asymmetric-dependence framing is itself becoming a tool of statecraft, independent of the underlying physical scarcity. Two CNAS studies71 one on mitigating U.S. dependence directly and one on the broader resource race72 reach a similar conclusion: stockpiling alone will not solve a problem that is fundamentally about manufacturing capacity, not inventory. The remedy is getting ahead of the problem: research into alternative materials and domestic component manufacturing capacity so that even a full Chinese mineral cutoff would leave the country without a buyer for the finished component.

Critical minerals remain important, but the AI compute race is shifting leverage toward critical components. For the West to compete with China on rare earth leverage, it must stay ahead as the chokepoint continues its climb up the supply chain. The Council on Foreign Relations' economic security task force73 and the U.S.-China Economic and Security Review Commission74 have both begun grouping minerals, semiconductors, and data-center components into a single chokepoint category for policy purposes, evidence that this shift is already shaping how Washington thinks about the problem. Should the West fail to match China in components, it risks a lasting disadvantage: a China insulated by its own dual circulation strategy, sealing off rare minerals from the world and forcing the West into dependency in the defining technological race of the decade.

Notes & References