The Mirror of Sanctions: China's Mineral Statecraft

The standard Western account of China's mineral dominance is a cost story. Cheap power, lax permitting, three decades of subsidy, so naturally the refineries ended up in one place. The account is true and almost entirely beside the point, because since 2023 China has run a sequence of export controls that reframes the same dominance as something else: a control story. The distinction is the whole argument. A cost advantage can be competed away by anyone willing to pay more or pollute more. Control has to be physically rebuilt somewhere else, and that takes time the West has not yet spent. Read the controls not as trade policy but as a series of live experiments in coercion, each one measuring how much pain Beijing can impose on a Western supply chain, how fast, and at what cost to itself. The early results are unambiguous: a great deal, almost immediately, and very little.

Dependence is priced in replication time, not market share

The variable that actually governs this problem is not China's share of any market today. It is the time the rest of the world would need to rebuild the single hardest step in the chain. For most critical minerals that step is not the mine. A deposit is capital plus permits, and with enough of both it can be brought into production in something like a decade, slow but finite, and buyable. The hardest step is separation and refining, and underneath it sits the part no balance sheet captures: tacit process knowledge. A solvent-extraction circuit that pulls one rare-earth element away from a chemically almost identical neighbour is tuned over years by engineers who learned on the circuit before it. That expertise was not in the patents China banned from export in December 2023. It lives in a workforce, and a workforce does not transfer on the schedule of a subsidy. This is why Western policy keeps missing the target. The instruments are denominated in dollars; the binding constraint is denominated in years.

A ladder, climbed in lockstep with Washington

Seen this way, the sequence of controls is not a list of grievances but a ladder, and each rung is chosen to maximise the adversary's adjustment cost per unit of revenue Beijing gives up. Gallium and germanium came first, in mid-2023, about a month after the United States and the Netherlands tightened semiconductor-equipment exports. They are close to an ideal weapon: China supplies roughly 98 percent of the world's gallium, the West almost none, and gallium is a minor byproduct of alumina refining, so the revenue forgone in restricting it is trivial. Graphite followed at the end of 2023, antimony in 2024. The day after Washington widened its chip controls in December 2024, China imposed its first mineral ban aimed at a single country, on gallium, germanium and antimony to the United States. Rare-earth magnets sit higher on the ladder, restricted in April 2025 in answer to tariffs, because withholding them costs China a genuine export industry, but they also bite hardest, since no Western magnet supply exists at scale. The selection is a revealed preference. Beijing reaches first for the materials where its own pain is least and the West's is greatest.

Why the weapon is rationed

The more revealing question is not why China uses these controls but why it keeps switching them off. The April 2025 rare-earth rules were eased within weeks. The sweeping extraterritorial regime of October 2025, which claimed licensing power over any product made anywhere using Chinese rare-earth content or process technology, was suspended under a trade truce almost as soon as it appeared. The familiar explanation is the lesson of 2010, when China throttled rare earths to Japan, lost the resulting case at the World Trade Organization in 2014, and watched the episode fund a decade of diversification. The deeper explanation is reflexive. Every actual use of the lever accelerates the diversification that destroys the lever. A threat left credible but unexecuted preserves the dependence indefinitely; a threat carried through finances its own replacement. The optimal policy for a monopolist who understands this is to demonstrate the capability, collect the concession, and stand down before the target's response hardens into capacity. The restraint is therefore not weakness. It is a supplier managing the half-life of its own monopoly, which is the more sobering reading, because it means Beijing is pricing the option, not bluffing with it.

The recipe is the chokepoint, and the West sold the recipe

The dominance that matters is in the recipe, not the rock. China mines perhaps 60 to 70 percent of the world's rare earths but refines around 90 percent of them, and it is a net importer of the ore it separates. The uncomfortable history is that the West did not so much lose this capability as offshore it on purpose, because separation chemistry is dirty and the regulatory cost of doing it at home rose faster than the strategic cost of not doing it. Mountain Pass in California, once the centre of the Western industry, spent years shipping its concentrate to China to be separated. The dependence is, in part, an environmental bill deferred for thirty years and now presented with interest. China's technology export bans, the 2023 prohibition on separation and magnet know-how and the 2025 rules that copy the logic of the American foreign direct product rule almost line for line, are an attempt to move the contest one level deeper: from controlling the material to controlling the method, and from the method to anyone else's supply chain that touches it.

Prices clear shortages where the constraint is capital. Where the constraint is accumulated process knowledge, a four-fold price buys smuggling and a handful of pilot plants, not an industry.

Price is not the constraint that clears this

The behaviour of prices under the controls disposes of the most comfortable Western assumption, that a shortage is a price signal and a price signal summons supply. Germanium roughly tripled after 2023. Antimony, long a sleepy metal around 11,000 to 13,000 dollars a tonne, reached almost 60,000 dollars by mid-2025. Chinese gallium and germanium export volumes fell by more than 90 percent in the months after licensing began. None of it produced a Western supply chain on the timescale that mattered. What it produced was a record price and a flow of metal that suddenly claimed origins that barely exist: nearly 3,800 tonnes of antimony oxide reached the United States from Thailand and Mexico in five months after the 2024 ban, from two countries that mine almost no antimony, which is to say relabelled Chinese material. The genuine new Western capacity that did appear, a first North American primary-gallium stream from a Quebec refinery, antimony projects pulled forward by the price, is real and years from scale. A price spike bought time, not supply.

What the response quietly admits

Read against that, the flagship Western responses are confessions. When the United States Department of Defense took a stake of roughly 15 percent in MP Materials in July 2025 and guaranteed a floor of 110 dollars a kilogram for its main rare-earth oxide, about double the prevailing price, it was not building an industry. It was guaranteeing a price because the market will not clear at the level China sets, which is the cost story revealed to have been downstream of the control story all along. The plant's near-term magnet target is around a thousand tonnes, under one percent of Chinese output. The European Union's Critical Raw Materials Act sets a 2030 goal of processing 40 percent of demand at home, the one benchmark its own auditors already expect it to miss. Even if every announced Western refining project arrives on schedule, ex-China refined output reaches perhaps a quarter of demand, and the heavy rare earths on China's 2025 list, dysprosium and terbium, stay almost entirely Chinese.

What would change this view

Desk discipline requires naming the evidence that would overturn the argument, because a thesis that cannot be falsified is not analysis. We would revise it if heavy rare-earth separation reached roughly a third of capacity outside China by 2030 at unsubsidised cost; if a Western magnet maker cleared automotive-grade volume without a government price floor beneath it; if the grey-market reroutes proved to be genuine new mine and refinery output rather than relabelled Chinese metal; or if China executed a full and sustained cutoff and absorbed the lost revenue, which would tell us it now values coercion above the half-life of its monopoly. None of these has happened. Until one of them does, the controls are not a passing aberration to be waited out. They are the visible operation of a structural lever.

What it reveals

The mirror is the lesson, not the metaphor. The West built its statecraft, the sanctions and entity lists and the foreign direct product rule, on the premise that leverage lives at the top of the value chain, in design and finance and intellectual property. China has shown that the bottom of the chain, the refined oxide and the separation recipe, holds leverage of the same order, and that whoever owns the step with the longest replication time owns the schedule. It is the same chokepoint Europe could not sanction in Russia and the United States could not reshore, seen now from the side of the party that holds it deliberately. The export controls did not create the dependence. They switched on the light in a room the West had been standing in, comfortably, for twenty years.