Not an embargo: China is using the pen to control rare-earth supply

Everyone is watching for a mineral blockade. The more powerful instrument is quieter, smarter and already in use — and India needs to learn how to survive it.

In May 2025, Ford stopped building the Explorer at its Chicago assembly plant for a week — not for want of chips, steel or workers, but for want of small rare-earth magnets used throughout the vehicle. They were stuck behind an exports-approval queue on the other side of the world.

India has substantial rare-earth resources in monazite, yet imports all its sintered magnets essentially because the rare earths come bundled with thorium. ([Photo credit: Reuters])

Nobody had declared an embargo. Beijing had simply asked the world to fill in forms. Deng Xiaoping’s widely cited 1992 remark — “The Middle East has oil. China has rare earths” — is usually read as a boast about geology. Three decades later, it looks more like an industrial-policy plan that worked.

Even though China holds less than half of the world’s rare-earth reserves, its control grows much stronger at every stage of production. According to the International Energy Agency (IEA), China accounts for about 60% of mining, 91% of refining, and 94% of finished-magnet manufacturing.

Ultimately, while the raw minerals are found worldwide, China overwhelmingly dominates the processing and magnet-making stages. The strategic resource, in other words, is not the rock. It is the ability to transform it — and increasingly, the paperwork governing access to it.

Why a licence beats an embargo

When Beijing restricted seven heavy rare earths and related magnets in April 2025, it did not ban their export. It required licences. Assembly lines paused across several countries. The IEA estimates that the full implementation of China’s subsequently expanded controls could put up to $6.5 trillion in annual downstream production outside China at risk.

The approvals tell the story: some shipments to the US have recovered, while Japan — after a diplomatic rupture with Beijing — had by July gone nine straight months without Chinese dysprosium oxide. Access is not denied. It is managed — by country, company and end-use.

Licensing works like a precision control valve rather than an outright ban, making it a far more powerful economic tool. Unlike a total trade ban, a licensing system can be dialled up for one nation and relaxed for another, keeping buyers hooked. Since customers still get occasional supplies, they find it harder to justify spending the money needed to build their own alternatives. On top of that, requiring companies to apply for permits reveals valuable inside information about who is buying what, how much they need, and what they are building. For instance, when supplies tightened in 2025, the manufacturer Bosch noted that its suppliers had to hand over detailed business and factory data just to get approved.

Export controls are beginning to resemble programmable sanctions for physical supply chains. Financial sanctions became powerful because they could target specific banks, companies and transactions. Mineral controls can bring similar granularity to industrial production.

An oil shock can stall an economy. A withheld magnet can stop a particular factory — and a licence desk can influence which one, and when.

China has tried the blunter instrument before. After a diplomatic dispute in 2010, rare-earth shipments to Japan were disrupted, and prices of some elements rose more than tenfold. Japan responded by financing alternative suppliers, reducing rare-earth use and accelerating substitution. Its dependence on Chinese rare earths fell from roughly 90% to about 60%.

The lesson was paradoxical: use a chokepoint too aggressively, and you finance its destruction.

Markets can route around chokepoints. The problem is that it can take a decade. The strategic advantage of the licence is that it can keep that decade from ever quite beginning.

The Mountain Pass lesson

Mountain Pass, the Californian mine that once supplied much of the world’s rare earths, illustrates another problem. Environmental issues led to the suspension of processing in 1998, and mining stopped in 2002. After the 2010 scare, Molycorp spent heavily to revive the operation. Rare-earth prices then collapsed, and by 2015, it was bankrupt.

Building alternative capacity is not enough if it must compete against a dominant incumbent capable of influencing prices. Washington has recognised this. In 2025, the US Department of Defence entered a partnership with MP Materials, which operates Mountain Pass, including a decade-long $110-per-kg price floor — roughly double the prevailing prices — for neodymium-praseodymium and commitments around its magnet output. Strategic mineral policy cannot end when the refinery opens. Someone must ensure it survives.

This is also why ‘critical minerals’ is a misleadingly convenient label. Copper primarily has a scale problem. The IEA still projects a substantial supply deficit by 2035. Rare earths have a concentration problem. Tiny volumes underpin trillions of dollars of downstream production. Battery minerals face volatility as prices and technologies change.

Geology may be sovereign. Processing technology, capital, and markets are often not. A country can own the ore while somebody else owns the refinery.

Sovereign AI begins in a refinery

The race for sovereign AI is usually discussed in terms of models, GPUs, data and electricity. Beneath that digital stack, however, lies a material one. Gallium powers advanced power electronics; germanium is important to optics and fibre.

China banned exports of both to the US in December 2024. Under the current trade truce, that restriction is suspended until November 27, 2026, with military end-users still excluded. The reprieve has an expiry date printed on it.

Rare-earth magnets are used in motors, actuators, hard drives and robots. AI data centres require enormous amounts of electrical infrastructure, including copper, aluminium and transformers. Countries trying to reduce dependence on foreign software and compute remain dependent on foreign atoms to build the machines on which that software runs.

The sovereign AI stack begins upstream of the data centre — in a refinery, and in a substation queue.

India’s rare-earth problem, & thorium opportunity

India has begun responding. The ₹16,300-crore National Critical Mineral Mission covers exploration, production, processing, recycling and overseas acquisition. KABIL has secured lithium exploration rights in Argentina.

More importantly, India is attacking the midstream gap. The government has committed ₹7,280 crore to create 6,000 tonnes a year of sintered rare-earth permanent-magnet capacity; its 2026 tender attracted bids from 20 companies.

But India’s most interesting problem lies deeper. The country has substantial rare-earth resources in monazite yet imports essentially all its sintered magnets. One reason is particularly Indian: the rare earths come bundled with thorium.

Because monazite contains thorium, the Indian government maintains strict legal control over it under the Atomic Energy Act. This creates a dilemma. The result is a collision between two strategic objectives: protecting the resource underlying India's long-term nuclear programme can complicate commercial exploitation of the rare earths needed today.

India’s thorium plans are finally gaining some traction. In April 2026, the 500-megawatt fast-breeder reactor at Kalpakkam reached a key technical milestone, bringing India a step closer to tapping its vast thorium reserves. Additionally, the December 2025 SHANTI Act opened nuclear energy to greater private participation, while NTPC and Clean Core have announced exploratory work on potential applications of ANEEL thorium-based fuel, subject to regulatory approval.

Commercial thorium power has not arrived. But this suggests a reform worth examining: separate thorium custody from rare-earth commerce. India could split the regulations: allow private companies to extract and sell the valuable rare earths from monazite under strict supervision, while requiring them to hand over all the thorium directly to the government.

The magnets and the reactors are, quite literally, in the same sand.

If India can eventually exploit thorium at scale, it gains something more valuable than another mineral to export: an indigenous energy resource less exposed to external fuel chokepoints. In an economy building AI infrastructure, semiconductor fabs and electrified industry, mineral security and energy security increasingly converge.

India’s real leverage is demand

India cannot out-mine or out-refine China this decade. It should not try to. Its advantage is demand.

India will become one of the world’s largest markets for magnets, batteries and related materials across EVs, wind power, electronics and defence. It can aggregate that demand into long-term offtake agreements that make alternative processing capacity bankable, both at home and in partner countries.

India needs to take three clear steps to secure its mineral supply. First, it needs to audit its supply chains to identify hidden chokepoints and build stockpiles of vital materials such as rare-earth magnets, gallium, and germanium. These reserves must be measured in months of survival, not total weight. New domestic refiners also need guaranteed buyers and price protections to stay alive.

Second, India needs to build technical expertise in this area. Government tenders can fund factories, but they cannot buy decades of practical experience. Refining rare earths is a specialised skill learned only over time.

Finally, India must start recycling and alternative technologies to fill the gaps in the interim. E-waste from old phones and batteries can serve as local sources of materials, while substitutes like sodium-ion batteries can steadily reduce dependence on foreign sources.

True mineral security is not about announcing new schemes or mineral discoveries. The real test is simple. How many unblockable supply routes does India have for every critical mineral it needs, and how many months can its factories run if a foreign official refuses to sign a licence?

The long-term view

The global race for critical minerals will gather momentum over the next decade. Governments will build emergency stockpiles, fund local refineries, guarantee minimum purchase prices, and tie access to minerals directly to foreign trade and defence partnerships. However, despite these efforts, China will maintain a substantial lead in rare earths. Because its real advantage is not just the raw minerals buried in its soil, but the decades of specialised technical know-how, refining expertise, and massive industrial scale it has built over time.

By the 2050s, current global chokepoints will change. Recycling, new battery and magnet chemistries, and AI-assisted materials discovery will take centre stage. They should reduce some existing dependencies and create new ones. For India, if its thorium programme eventually succeeds at scale, it could remove another external vulnerability from an increasingly electrified economy.

By 2076, today’s list of ‘critical minerals’ may look as dated as a 19th-century strategic list built around guano, saltpetre and natural rubber. What will remain critical is capability: the ability to refine, substitute, recycle, innovate — and maintain enough alternatives to refuse coercion.

Deng saw the rocks coming. The sharper eye watches the paperwork.

The next geopolitical shock may not arrive with tankers stranded in a strait. It may look like that Ford factory: an assembly line stopped, a form pending, and a licence that never quite arrived.

(The views expressed in this article are personal)