Why protectionism cannot close the critical minerals gap, and how refining, industrial investment & supply-chain diversification reshape global power.
An average that folds every refinery into one number hides the stage at which dependence actually forms, so the chain has to be separated into reserves, mining, refining, intermediate materials and finished products. Mining is slowly diversifying. The IEA's full report finds that the top three producing countries supply around 76% of global mine output, projected to ease to 73% by 2035. Refining is "considerably starker": China holds just under 50% of global copper refining capacity, 70% of lithium refining, 75% of cobalt refining, 85% of magnet rare earth separation and over 90% of battery-grade graphite production. Rare earths, a family of distinct elements with separate markets, can be controlled one element at a time.
Placing the IEA's mining and processing figures for six minerals side by side reveals the distance between the two stages, and how differently each is expected to move by 2035.
| Mineral | Mining picture (IEA, 2025) | China's share of refining or processing (IEA, 2025) | IEA base-case direction to 2035 |
|---|---|---|---|
| Lithium | Australia, Chile and China together about 75% | 70% of refining | Top-three mining share falls below 70% |
| Graphite | China about 80% of mined natural graphite | 94% of battery-grade supply | China's mined share 53%; battery-grade share 91% |
| Magnet rare earths | China around 60% of mined supply | 85% of separation, down from 90% in 2024 | Mining 55%; separation 73% |
| Cobalt | Democratic Republic of the Congo (DRC) under export quotas since October 2025 | 75% of refining | Supply gap above 25% in 2035 |
| Nickel | Indonesia's share rising above 70% | Indonesia is the top refiner | Concentration keeps increasing |
| Copper | Latin America about 40% of mining | Just under 50% of refining capacity | Supply deficit of about 25% in 2035 |
If the table locates concentration in processing, the obvious follow-up is why a refinery cannot simply be built elsewhere once tariffs make imports dearer. The IEA's answer is that refining capacity "remains closely linked to established chemical processing infrastructure, accumulated technical expertise and proximity to downstream manufacturing." Capital costs for refining projects are 20% to over 150% higher outside the dominant supplier. That ecosystem of equipment, chemists, supplier density and scale does not respond quickly to import prices. A process proven in a pilot plant is not yet commercial supply: it must deliver volume, quality and consistency at competitive cost, and buyers must qualify the material first, one of the switching costs the IEA lists with testing and certification. Rare earth equipment made outside China can cost 5 to 12 times more and take two to three times longer to arrive, from manufacturers that "have not secured long-term established customers."
Changing the battery recipe looks like a way around the refinery problem, but the IEA's chemistry data show that it relocates exposure without removing it. Lithium iron phosphate (LFP) accounted for 55% of global electric vehicle (EV) batteries sold in 2025, and LFP packs were more than 40% cheaper on average per kilowatt-hour than lithium nickel manganese cobalt oxide (NMC) packs. Choosing LFP sheds the need for cobalt and nickel, but both chemistries still need lithium and graphite anodes, and China produced 98% of LFP cathode materials in 2025, two-thirds of nickel-based cathode material and more than 90% of anode material. Its October 2025 battery controls, now suspended, named LFP cathode materials among their targets.
Neither a new refinery nor a new chemistry escapes China's processing base quickly, and the export-control notices issued in Beijing since late 2024 showed what that concentration can be used for. The Ministry of Commerce (MOFCOM; 中华人民共和国商务部) issued Announcement No. 46 of 2024 on December 3, 2024, banning dual-use exports to US military users and stating that gallium, germanium, antimony and superhard materials would in principle not be licensed for the United States. On April 4, 2025, MOFCOM and the General Administration of Customs placed seven medium and heavy rare earths and related magnets under export licensing through Announcement No. 18. The European Central Bank (ECB) recorded that Chinese shipments of rare earth magnets fell by approximately 75% year on year in May 2025, and some carmakers paused production. In October 2025, Announcement No. 61 reached abroad, requiring foreign exporters to obtain a Chinese license for goods containing 0.1% or more by value of controlled Chinese-origin rare earth material.
Much then depends on reading these instruments precisely, since a licensing requirement is a gate that can be opened, slowed or shut. As of October 26, 2026, No. 18 remains in force. Announcement No. 70 suspended No. 61 and five related notices until November 10, 2026, and Announcement No. 72 suspended the US-specific gallium, germanium and antimony clause of No. 46 until November 27, 2026, leaving its military end-use ban intact. Beijing has since opened a reporting system for violations of strategic-mineral export rules, including transshipment through third countries. In September 2026, the two governments agreed to extend the Kuala Lumpur arrangement to January 10, 2027, but MOFCOM's announcement index shows no notice yet amending the November 10 date.
The legal record establishes capacity; motive is a separate question, and here Beijing's stated rationale and outside readings diverge. No. 18 cites the need "为维护国家安全和利益、履行防扩散等国际义务," which translates as safeguarding national security and interests and fulfilling international obligations such as non-proliferation. ECB economists judged the April measures to have been "introduced in retaliation for increased US tariffs on Chinese goods" and concluded that China "has shown its willingness to use these elements to pressure trading partners." The IEA recalls that China's 2010 quotas and licensing sent rare earth prices temporarily up by as much as ten times before they were ruled non-compliant with World Trade Organization (WTO) rules. Both readings can hold at once, since non-proliferation concerns are real for dual-use materials. The record does not show a permanent embargo; licensing, suspension and negotiated extension have been used as dials.
Export restrictions are often described as a Chinese instrument, yet several countries that mine what China refines use their own versions to capture more of the value chain. Indonesia stopped exports of nickel ore below 1.7% nickel content from January 1, 2020. The DRC's strategic minerals regulator replaced a cobalt export suspension with quotas from October 16, 2025, set at 7,250 metric tons for each of November and December and reviewable according to each firm's contribution to local value chains; the IEA attributes much of a roughly 130% rise in cobalt prices to those restrictions. Zimbabwe, which halted lithium concentrate shipments in February 2026, has reaffirmed a full ban from January 2027. The European Union (EU) has signaled restrictions on exports of permanent-magnet scrap, still at the proposal stage. Argentina shows the opposite model, open exports plus investment incentives, approving lithium expansions such as a 9,500-metric-ton-a-year addition at the Fenix project under its Large Investment Incentive Regime (RIGI).
Where producer countries restrict exports to capture more value from their resources, consumer economies have relied more heavily on trade measures to reduce their exposure to concentrated supply. The US record illustrates both the reach and the limits of that response. The Office of the United States Trade Representative (USTR) raised Section 301 tariffs on Chinese natural graphite and permanent magnets to 25% from 2026. In January 2026, Proclamation 11001 under Section 232 imposed no new tariff. It directed negotiations, told officials to "consider price floors for trade in critical minerals and other trade-restricting measures," and stated the diagnosis plainly: "Mining a mineral domestically does not safeguard the national security of the United States if the United States remains dependent on a foreign country for the processing of that mineral." A plurilateral agreement built on border-adjusted price floors remains a proposal without a released text.
Washington's own diagnosis points to processing, which raises the question of whether capital to build it is arriving at the pace the tariffs assume. Critical mineral investment declined by 9% in 2025, and battery-metal capital spending fell by more than 20%. Public finance commitments in advanced economies reached around US$65 billion, yet the IEA sees "a considerable gap" between commitments and disbursements. The pipelines are lopsided too: diversified rare earth refining capacity equals around two-thirds of expected mined supply by 2035, planned magnet output only one-third. A tariff reprices an imported magnet on day one; a plant still on the drawing board cannot answer that signal.
When tariffs arrive before capacity, users bear the cost, and how heavily depends on the minerals' share of final prices, where the IEA narrows the inflation worry. Critical minerals account for around one-quarter of battery cell costs but only about 3% of the price of an average EV. Tripling rare earth prices would add about 0.1% to the cost of a car; tripling battery material prices would lift EV and storage prices by around 5%. Tariffs on such inputs tend to produce a one-time shift in the price level, and USTR's four-year review cited analyses finding minimal impacts on economy-wide prices and employment. ECB research also found that pass-through from intermediate to consumer goods prices usually takes more than one year and can be partly absorbed in margins. Sustained inflation would require restrictions that keep tightening.
A small consumer-price effect does not make the policy cheap, because the larger costs lie in transition and competitiveness. Beyond the tariff and the capital-cost gaps already described, the competitiveness cost is the hardest to see: in Europe, gallium, dysprosium and terbium cost around five times Chinese domestic prices, a handicap for manufacturers competing with firms that buy at home. The gravest exposure concerns quantity. Full rare earth controls would put US$6.5 trillion a year of downstream production outside China at risk, around 7% of those countries' combined gross domestic product (GDP).
Critics of state intervention have to reckon with the rare earth exception noted at the outset, because it suggests that targeted support can move a market. The IEA credits new projects in the United States and higher output in Malaysia, "highlighting the role of targeted policy and investment support in enabling diversification." The clearest instrument is the July 2025 partnership between MP Materials and the Department of Defense, which set a 10-year price floor of US$110 per kilogram for neodymium-praseodymium, guaranteed offtake from a magnet plant due to begin commissioning in 2028, alongside federal equity and a loan. Japan points the same way. Its state-backed partnership with Lynas secured up to 65% of the Australian producer's dysprosium and terbium for Japan, and the IEA reports Japanese rare earth demand nearly 30% below 2010 levels after years of measures to reduce use.
That evidence supports instruments aimed at specific bottlenecks; it does not extend to blanket tariffs, which can strengthen the advantage they target. A duty on Chinese magnets, imposed while domestic magnet capacity is years away, raises costs for US motor and vehicle makers while Chinese competitors keep buying at home. Policies also pull against each other when governments subsidize processing while taxing the intermediates their manufacturers still need. Tariffs can also produce a misleading scoreboard: a country can assemble EVs or batteries at home while importing the refined lithium, cathodes and anode material inside them, so domestic assembly share is an incomplete measure of autonomy. Whether this entrenches China's position cannot yet be proven, though concentration has risen in most refining segments despite two years of tariffs and subsidies.
Targeted support still has to be designed so that it builds an industry and does not merely shelter one. Infant-industry protection earns its cost when it has an end date and benchmarks for output, cost and quality; open-ended support keeps uncompetitive plants alive at public expense. The IEA advises balancing incentives with fiscal exposure "through competitive tenders and performance-based eligibility," and the contracts for difference it describes require operators to repay the government when market prices exceed the strike price. Consistency matters as much, because a processing plant takes years to plan, permit and qualify, and a sudden change in tariffs, credits or eligibility can strand it midway. In July 2025 the One Big Beautiful Bill Act ended tax credits for new and used electric car purchases after September 2025; the IEA reports that US sales fell significantly in the last quarter and lists policy uncertainty among the reasons battery-metal investment retreated.
Even well-designed national support cannot efficiently reproduce every stage of a critical-mineral supply chain within a single economy. The next challenge is therefore to coordinate investment, processing capacity and supply agreements across allies and producer partners. The EU's RESourceEU plan mobilizes about €3 billion over 12 months and the Critical Raw Materials Act sets 2030 benchmarks of 10% domestic extraction, 40% processing, 25% recycling and no more than 65% from any single third country. Australia legislated an A$1.2 billion Critical Minerals Strategic Reserve in April 2026. Canada's Budget 2025 proposed C$2 billion over five years for what is now the Canada Critical Minerals Accelerator. South Korea, which holds strategic mineral stockpiles, added US$163 million to its Supply Chain Stabilization Fund in 2025; it leads outside China in battery material recovery, with 6% of global capacity, and Korea Zinc's planned refinery of 54 metric tons of germanium and 44 metric tons of gallium a year could double diversified germanium supply. India approved a ₹7,280 crore magnet scheme for 6,000 metric tons a year, about US$767 million at the Financial Benchmarks India (FBIL) rate of ₹94.9432 per dollar on October 23, 2026 (one crore is 10 million), and a National Critical Mineral Mission of ₹34,300 crore, about US$3.6 billion. Group of Seven (G7) leaders agreed in June 2026 to cut reliance on any single non-G7 supplier of rare earths and magnets to below 60% by 2030.
Diversification plans are usually written as if China were only a rival supplier, though the same evidence shows it as investor, customer and negotiating counterpart. Chinese companies, including Sinomine, Yahua and Huayou, are building or running lithium sulfate plants in Zimbabwe to meet that country's processing deadline. In October 2026, MOFCOM said it would keep facilitating rare earth export license approvals for the EU. China also guards its own supply, having curbed sulfuric acid exports in May 2026 amid Middle East disruptions. A strategy aimed at single-supplier exposure, the G7's own formulation, leaves room for continued trade with China; one aimed at exclusion would forfeit the system's largest processor and a major investor and customer.
How governments combine these tools will shape the future of the critical-minerals market, influencing where industrial capacity develops and how much the energy transition costs. Under rival blocs, export controls multiply, governments tighten local-content requirements, and competing industrial subsidies encourage firms to duplicate production across politically aligned markets. Such fragmentation could reinforce supply bottlenecks, raise manufacturing costs and prolong the substantial price disadvantages European manufacturers face for certain critical minerals. Under negotiated diversification, more predictable export-licensing arrangements, long-term offtake contracts and investment in processing capacity across producer countries could gradually reduce dependence on individual suppliers. This approach would not eliminate China's role in global supply chains, but it could give buyers more alternatives and make disruptions less damaging. Under substitution, advances in rare-earth-free motors, alternative battery chemistries and recycling could reduce demand for particular critical minerals or recover more material from existing products. The strategic importance of individual refiners could consequently diminish, although the pace of change would depend on commercial viability, manufacturing scale and the time required to qualify new materials and technologies. These scenarios are not forecasts; they illustrate how policy and technological choices could reshape supply security, industrial competitiveness and geopolitical leverage.
Apple News, Google News, Feedly, Flipboard, and WhatsApp Channel
