Why the strategic bottleneck of the energy transition lies not in the ground but in the furnace, and why that distinction has turned obscure metals into levers of statecraft

I. The fault line: reserves are abundant, refining is not

The reflexive Western anxiety about critical minerals frames them as a scarcity problem, a race for deposits in contested geographies. The data tell a different story. The constraint is not the mine; it is the plant. According to the International Energy Agency's Global Critical Minerals Outlook 2025, China is the dominant refiner for 19 of the 20 strategic minerals it analysed, holding an average market share of around 70 percent ^[1]. That figure describes the midstream, the unglamorous chemistry of leaching, separation, calcination and purification that turns ore into battery-grade and magnet-grade material, not the mines themselves.

The asymmetry is stark when extraction and processing are placed side by side. China processes roughly 60 percent of the world's lithium and cobalt and controls over 90 percent of refining capacity for both graphite and rare earth elements, even though it mines only a fraction of the underlying ore ^[1]. Lithium is pulled from Australian hard rock and South American brine; cobalt comes overwhelmingly out of the Congolese Copperbelt; yet a disproportionate share of both passes through Chinese furnaces before reaching a cathode or a magnet. The geology is distributed. The chemistry is concentrated.

Lithium makes the point in miniature. World lithium mine production reached a record of roughly 240,000 tonnes in 2024, with Australia, Chile and China together accounting for more than 85 percent of mine output ^[15]. But mined lithium is not usable lithium: it must be converted to battery-grade carbonate or hydroxide, and that conversion step, not the spodumene mine or the brine pond, is where China's share dominates ^[1]. A country can hold world-class deposits, as Australia does, and still depend on a rival's refineries to monetise them. Reserves confer potential; refining confers control.

The world does not have a mineral shortage. It has a refining monopoly, and a monopoly is a policy, not an accident.

This concentration is not loosening; it is tightening. The IEA finds that the average market share of the top three refining nations for key energy minerals rose from around 82 percent in 2020 to 86 percent in 2024, and is projected to decline only marginally, to about 82 percent, by 2035 on current trends ^[1]. Some 90 percent of refined-supply growth between 2020 and 2024 came from the single top supplier in each case: Indonesia for nickel, China for cobalt, graphite and rare earths ^[1]. Diversification, in other words, is not happening at anything like the pace the rhetoric implies.

II. The Congolese paradox: African ore, Chinese metal

Nowhere is the divorce between resource and refining more visible than in cobalt. The Democratic Republic of the Congo accounted for an estimated 76 percent of world cobalt mine production in 2024, far ahead of Indonesia at around 10 percent, according to the United States Geological Survey ^[2]. The DRC is, by any measure, the indispensable supplier of a metal still central to high-density batteries.

And yet the value, and the leverage, sit elsewhere. Chinese companies are estimated to control around 80 percent of the Congolese cobalt market, with stakes in 15 of the largest copper-cobalt mines in the country ^[3]. CMOC Group, operator of the Tenke Fungurume and Kisanfu mines, reported mined cobalt output of roughly 114,000 tonnes in 2024, making it the world's largest cobalt miner by volume and displacing Switzerland's Glencore from the top spot ^[3]. The ore is African; the corporate control, and almost all of the downstream refining, is Chinese.

Africa supplies the metal of the energy transition and captures the least of its value. That is not a market outcome. It is a structural choice that can be unmade.

Kinshasa has begun to test its own leverage. Faced with a price collapse, cobalt fell from a 2022 peak near USD 82,000 per tonne to roughly USD 20,000 by February 2025 amid oversupply, the DRC imposed an export ban in February 2025, extended it, and in October 2025 replaced it with a quota system capping exports at around 96,600 tonnes per year for 2026 and 2027 ^[4]. The intervention pushed prices sharply higher ^[4]. It is a reminder that producer states are no longer passive, but also that controlling the mine is not the same as controlling the metal.

III. From dominance to coercion: the 2023–2025 export-control sequence

A dominant position becomes an instrument only when it is used. Between mid-2023 and 2025, Beijing crossed that line in a deliberate, escalating sequence, and the trigger each time was a US technology measure.

The opening move came on 3 July 2023, when China's Ministry of Commerce announced export controls on gallium and germanium, effective 1 August 2023, requiring licences for a wide range of compounds ^[5]. The timing was not coincidental: it followed tightening US semiconductor controls. China produces roughly 98 percent of the world's low-purity gallium and around 60 percent of germanium, per USGS data, so the controls landed on chokepoints where buyers had no quick alternative ^[6]. In October 2023, Beijing extended licensing to high-purity and high-quality graphite, effective December 2023 ^[7][1], graphite being a material in which China accounts for some 77 percent of natural and over 95 percent of synthetic production ^[7].

The escalation sharpened in 2024. After the US expanded chip-equipment restrictions on 2 December 2024, Beijing responded the next day, on 3 December 2024, by banning outright the export to the United States of gallium, germanium, antimony and superhard materials, the first time Chinese mineral controls were targeted at a single country rather than applied globally ^[7]. The antimony case is instructive on precision: following China's antimony restrictions (announced August 2024, effective September 2024), Chinese antimony shipments fell by roughly 97 percent and prices rose around 200 percent ^[7]. China holds about 48 percent of global antimony production and supplies 63 percent of US antimony imports ^[7].

IV. The rare-earth squeeze and the magnet monopoly

The most consequential turn came on 4 April 2025, when the Ministry of Commerce imposed licensing controls on seven medium and heavy rare earth elements, samarium, gadolinium, terbium, dysprosium, lutetium, scandium and yttrium, and on the permanent magnets made from them, in retaliation for US tariff increases ^[8][9]. These are precisely the elements that confer high-temperature performance on the NdFeB magnets at the heart of electric-vehicle motors, wind turbines and guided weapons.

Here the monopoly is near-total. China controls roughly 90 percent of global rare-earth processing and around 85–90 percent of NdFeB magnet production, and is the only country with processing capability at every stage of the magnet supply chain ^[10]. The effect was immediate: export volumes of magnets fell sharply in April and May 2025, and carmakers in the US, Europe and elsewhere were forced to cut utilisation or idle lines for want of magnets ^[9][1].

A single licensing decision in Beijing can stop an assembly line in Stuttgart or Detroit within weeks. That is the definition of strategic leverage, and it was demonstrated, not theorised, in 2025.

The de-escalation that followed should not be misread as resolution. After the Xi–Trump meeting, China in November 2025 suspended the additional rare-earth controls it had announced on 9 October 2025, for one year, and agreed to issue general licences for shipments to the US ^[11][12]. But, and this is decisive, the suspension did not lift the April 2025 controls on the seven elements and their derivatives; exporters of those items still require MOFCOM licences ^[12]. The valve was loosened, not removed. The architecture of control remains fully in place, available to be re-tightened.

V. The Western response: targets without tonnes, incentives without midstream

The policy reaction in Brussels and Washington has been real but structurally mismatched to the diagnosis. The European Union's Critical Raw Materials Act (Regulation (EU) 2024/1252), in force from 2024, sets 2030 benchmarks for the EU's annual consumption: at least 10 percent extracted domestically, at least 40 percent processed, at least 25 percent recycled, and, crucially, no more than 65 percent of any strategic raw material at any processing stage sourced from a single third country ^[13]. The 40 percent processing target and the 65 percent cap are the right instruments in principle: they aim squarely at the refining chokepoint rather than at mining alone. The Act also offers Strategic Projects faster permitting, 27 months for extraction, 15 for processing ^[13].

The United States took a different route through the Inflation Reduction Act, conditioning the consumer EV tax credit (Section 30D) on critical-minerals sourcing: from 2024, a rising applicable percentage of mineral value must be extracted or processed in the US or a free-trade-agreement partner, and from 2025 minerals cannot be processed by a "foreign entity of concern", a category that captures Chinese-controlled firms, defined down to a 25 percent ownership or board threshold ^[14]. This is industrial policy by tax code, and it has catalysed investment. But both instruments share a weakness: targets and incentives do not by themselves build refineries, which are capital-intensive, slow to permit, environmentally contentious and, critically, must compete against incumbent Chinese capacity that can be priced strategically.

VI. Institut Vidocq assessment, the chokepoint is the chemistry

The Institut's reading of the 2023–2025 sequence is that the prevailing Western mental model is wrong in a way that matters. The contest is framed as a scramble for deposits; it is in fact a contest over midstream industrial capacity. Three implications follow.

First, resource diplomacy without refining capacity is hollow. Securing offtake from a Congolese mine or an Australian spodumene operation does little if the only economic path to finished material runs through Chinese plants. The DRC's 76 percent of mined cobalt ^[2] coexists with Chinese control of roughly 80 percent of that same cobalt's commercial chain ^[3]; ore access and metal access are different securities.

Second, the coercive instrument is calibrated, reversible and asymmetric. The 97 percent collapse in antimony exports ^[7] and the 2025 magnet shortages ^[9] show that licensing regimes are not blunt embargoes but fine-grained valves, turned by country, by element, by end-use, and just as easily turned back, as the November 2025 partial suspension demonstrated ^[11][12]. Reversibility is itself the weapon: it keeps the threat live and the dependent party disciplined.

Third, for Africa, the strategic question is value capture, not merely extraction. The continent that supplies the defining metal of electrification captures almost none of its midstream rents. The Institut's view is that the durable answer, for Europe's security and Africa's development alike, lies in co-located refining: building separation and processing capacity at or near the resource, under transparent governance, rather than shipping concentrate to be transformed elsewhere. Kinshasa's quota experiment ^[4] is a crude first step toward repricing leverage; a refining strategy would be a structural one.

VII. Conclusion, controlling the valve, not the vein

The decade's defining mineral question is not whether the world holds enough cobalt, lithium or neodymium. It demonstrably does. The question is who controls the furnaces, the solvent-extraction trains and the magnet sintering lines that stand between an ore body and a working machine. Between 2023 and 2025, Beijing answered that question by turning a refining monopoly into a coercive instrument, gallium, germanium, graphite, antimony and the heavy rare earths deployed in sequence, each time as retaliation, each time with surgical effect.

The Western response has correctly identified the chokepoint on paper, the EU's 40 percent processing benchmark and 65 percent supplier cap, the US foreign-entity-of-concern rules, but has not yet built the tonnes. Until refineries exist outside China at scale, and until producing states in Africa and beyond capture the midstream rather than exporting raw concentrate, the valve will remain in one set of hands. Reserves are a map of where power could be. Refining is where power is. The strategic task of the coming decade is to move the second to match the first.