Why the world keeps coming back to China for its metals

China has squeezed exports, expanded supply and lived through violent price swings. Foreign buyers have felt the shortages. Rival producers have felt what happens when the material gets cheap again.

Research cut-off: 29 September 2026. Annual figures retain their stated year.

Japanese manufacturers got a lesson in mineral dependence in 2010. After a fishing-boat collision near the disputed Senkaku/Diaoyu islands and Japan's detention of the Chinese captain, their rare-earth shipments from China ran into severe delays.

Japan's trade ministry later documented the slowdown in customs clearance and its gradual resolution after about two months. China denied imposing an embargo. It was a disputed political episode with a very practical consequence: buyers couldn't count on their normal deliveries. ^[18]

During that controversy, China's premier gave a public assurance:

“China will not use rare earth as a bargaining chip.”

Wen Jiabao, Chinese premier, China–EU Business Summit, Brussels, 6 October 2010. ^[19]

The written record of China's wider export restrictions is less ambiguous. Quotas had been cut sharply. Prices soared in 2010–11. New suppliers attracted money, including the business rebuilding America's Mountain Pass mine. Then prices fell, and the challenge changed from finding material to selling it for enough to stay in business.

That second half of the story matters just as much. A shortage can make a rival mine look indispensable. Cheap supply can make the same mine look unfinanceable. China's industrial scale sits behind both problems, although the evidence does not justify treating every price movement as an order from Beijing.

01 / Actual rare earths: China built beyond the mine

The huge excavations at Bayan Obo, in Inner Mongolia, give China's mineral story a physical scale that a trade announcement can't. The satellite photograph above is real, although it is an archival, false-colour image from 2006. It shows a mining landscape, not a map of where individual rare-earth elements are concentrated.

China's geology is substantial. USGS reports 44 million tonnes of rare-earth reserves, expressed in rare-earth-oxide equivalent, and estimates 270,000 tonnes of mine production in 2025. A reserve is material considered economically recoverable under the estimate's assumptions. It isn't all the rare earths in the ground, and it certainly isn't a warehouse of finished magnets. ^[3]

There are different geological stories within the country. Northern deposits such as those around Baotou are associated particularly with light rare earths. Southern ion-adsorption deposits, including those in Jiangxi, helped supply medium and heavy rare earths. The mix matters: a large quantity of one element cannot automatically replace a smaller quantity of another. ^[14]

Neodymium and praseodymium are central to the strongest widely used rare-earth magnets. Dysprosium and terbium can help magnets retain performance at higher temperatures. Getting the right ingredients into a reliable product requires much more than digging up a mixture that contains them.

2010–11: the export squeeze was real

The Japan dispute took place against a much broader cut in permitted exports. China reduced its full-year rare-earth export quota by 37.1% in 2010. The second-half allocation alone was 70% below the equivalent period in 2009. Those are different comparisons; the often-repeated 70% figure was not a cut to the whole year's quota. ^[18]

China presented restrictions as part of conserving resources and tackling environmental damage. Foreign manufacturers saw a supplier limiting access to inputs they needed. The US, EU and Japan brought a WTO case in 2012, and the findings adopted in 2014 went against China. The WTO's account describes restrictions that gave Chinese manufacturers preferential access to material. ^[13]

That advantage extended beyond a miner receiving a better price. A manufacturer outside China had to obtain restricted exports; one operating inside China had access to the domestic market. Keeping more of the material at home could help pull the next stage of manufacturing there too.

The price shock was large enough to change investment decisions. In its later filings, Molycorp reported a 2011 annual average FOB China price of $251 per kilogram for neodymium oxide. By 2014, the same reported measure was $65. Anyone financing a plant around the first number had to survive the second. ^[21]

2015: more exports, lower prices, an American bankruptcy

China removed the challenged export quotas and duties in 2015. Its exports of rare-earth materials reached 26,800 tonnes in January–October, up 20% on the same ten months of 2014, according to USGS. More material was reaching foreign buyers, in a market already struggling with excess supply. ^{[13, 20]}

USGS year-end quotations for neodymium oxide of at least 99.5% purity fell from $190–200/kg in 2011 to $37–42/kg in 2015. ^[20]

Molycorp filed for Chapter 11 protection on 25 June 2015. Its own account cited weak prices, uneven demand and delays bringing Mountain Pass up to the intended operating level. The business had not reached break-even operating cash flow even before interest. Cheap material was a brutal problem for a plant that was still struggling to deliver its promised economics. ^[21]

The tempting version is that China turned off supply, waited for competitors to build, then flooded the market to kill them. Parts of that sequence are documented. The export restrictions were real, exports subsequently grew, and collapsing prices hurt the American challenger. A centrally directed plan to bankrupt Molycorp is a stronger claim than these records establish.

There were other sources of pressure. USGS reported illegal Chinese production despite government efforts to contain it, excess inventories, and rising output from the Australian–Malaysian supply chain. Some demand was changing too: the move towards LED lighting reduced demand for rare earths used in phosphors. Its 2015 report described China adding to state stockpiles, which hardly fits a story in which every cheap tonne came from an official stockpile dump. ^{[20, 24]}

Winning a trade case had not guaranteed a profitable American mine.

Cheap wasn't always what Beijing wanted either

China's producers also compete with one another. That matters when outsiders imagine one perfectly coordinated seller choosing the world price.

In March 2021, industry minister Xiao Yaqing complained about exactly this:

“Our rare earth does not sell at a "rare" price, but at an "earth" price.”

Xiao Yaqing, China's minister of industry and information technology, 1 March 2021; official English transcript. ^[23]

He blamed competition and price-cutting, while also discussing wasteful extraction and environmental complaints. His words describe a government trying to improve the returns from an industry that could undermine its own margins. They are not evidence that China lacks power over supply. They show why the interests of the state, a miner and a magnet factory don't always line up. ^[23]

2022–24: more supply, another fall in prices

The price problem returned. China's final 2023 mining allowance was 255,000 tonnes of rare-earth-oxide equivalent; the first two 2024 allocations totalled 270,000 tonnes. These were production permissions, not a record of how much every mine actually delivered. The IEA described 2023 as a year of Chinese oversupply and weaker-than-expected demand from downstream manufacturers. ^[22]

USGS's annual average FOB price for combined neodymium-praseodymium oxide fell from $124/kg in 2022 to $55 in 2024, a drop of about 56%. This is a different product and price measure from the historical chart above. Its estimated 2025 average recovered to $69, still well below 2022. ^[22]

For a new foreign producer, a price recovery that may arrive next year does not pay this year's bills. The established industry can keep expanding while a smaller rival delays a project, seeks more money or stops work. Chinese producers also suffer when prices fall; the IEA reported a sharp decline in the profits expected by China Northern Rare Earth for 2023. ^[22]

That is the uncomfortable part of competing with such a large supplier. Countries want an alternative ready when exports tighten, but the company building it has to survive the years when material is plentiful. Our investigation of US mineral dependence and processing gaps follows the price guarantees and purchasing commitments intended to bridge that gap.

The work between the rock and the motor

Ore first has to be mined and upgraded. Chemical processing then separates elements that arrive together. The resulting oxides need further conversion into metals and alloys before a magnet manufacturer can use them.

These aren't interchangeable workshops. A company that makes an acceptable concentrate has not necessarily learned to separate dysprosium, produce clean metal or deliver a magnet that passes an automotive customer's tests. Each new operation needs equipment, people who know how to run it, and enough orders to keep it working.

The IEA's comparable 2024 figures show how far China carried that process: about 60% of magnet rare-earth mining, 91% of refining and 94% of sintered permanent-magnet production. The mining figure covers the four principal magnet elements, not the whole rare-earth basket counted by USGS. ^[4]

A mine also produces a mixture dictated by its geology. The valuable magnet ingredients can arrive alongside much more abundant cerium and lanthanum. Selling the sought-after elements doesn't automatically give the operator a profitable market for everything else it extracts. The IEA identifies that mismatch as another obstacle to financing new supply. ^[4]

That changes how to read a resource announcement. A very large total can still leave a project short of the element a customer needs, or burden it with products few buyers want. The useful number is the amount of each saleable material the plant expects to recover, at a cost customers will support.

For a new overseas mine, selling concentrate into that established system can be the sensible commercial choice. For the government financing the mine to reduce dependence, it may be an uncomfortable result. More non-Chinese ore can still mean more business for Chinese processors.

Our briefing on China’s rare-earth mines, production and processing capacity identifies the operating sites behind these figures.

The industry wasn't born fully formed

China's own 2012 white paper traces the industry's development back to the 1950s, with faster growth after the reforms of the late 1970s. It described research, equipment production and materials manufacturing alongside extraction, but admitted weaknesses in higher-value products. Large processing capacity had not automatically made China good at every finished component. ^[14]

The subsequent magnet figures show how much changed. China made roughly half the world's sintered permanent magnets in 2005. By 2024, the IEA puts its share at 94%. A large domestic manufacturing base gave suppliers orders, experience and reasons to improve. Customers and processors grew together. ^[4]

There was a cost behind the cheap material. The 2012 white paper acknowledged polluted water, damaged farmland and radioactive residues from parts of the industry. Those burdens helped make environmental reform necessary. They also make the old low prices a poor guide to what a new, responsibly operated plant elsewhere should cost. ^[14]

China's lead combines real technical development with a history of costs that weren't always fully reflected in the product. Explaining it solely through cheap labour misses the chemistry. Explaining it solely through brilliant planning misses the waste.

02 / Strategic metals, by EarthRarest's definition

Our strategic-metals group contains antimony, bismuth, gallium, germanium, hafnium, indium, rhenium and tellurium. This is EarthRarest's editorial definition, not an official eight-metal list shared by every government. None of the eight is a rare-earth element.

They deserve a separate section because many are produced almost as a second business inside the first. A plant earning its living from aluminium, copper or zinc can recover another element along the way. Whether it does so depends on the feed, the equipment and the price someone will pay.

That small decision, repeated across a very large industry, can leave the rest of the world dependent on a surprisingly small set of suppliers.

Gallium: the aluminium business has a second product

Gallium is the clearest example. USGS estimates that China supplied 99% of primary, low-purity gallium in 2025. The qualification matters: this is the first recovered metal, before the further purification and manufacturing needed for electronic materials. It does not mean China made 99% of every gallium semiconductor. ^[2]

Most primary gallium comes from processing bauxite, the ore used to make alumina and then aluminium. The gallium is present in small amounts. A suitable refinery can install equipment to recover it from its process streams instead of letting it pass through without becoming a product. Zinc processing offers another route. ^{[2, 5]}

Think about the investment decision. An existing alumina refinery already handles large volumes of ore and pays for much of the infrastructure. Adding gallium recovery can share that foundation. Starting a bauxite mine and an alumina business solely to obtain gallium would be a very different proposition.

China's enormous aluminium industry therefore offers more than a large domestic market. It provides places where recovery can happen. A mineral that sounds like a stand-alone mining problem turns out to depend on the economics of another metal.

The distinction shows up in the American industry too. USGS's 2026 report recorded no US primary gallium recovery since 1987, while describing a New York company refining imported low-purity gallium and new scrap. The country could perform a valuable later step while still depending on somebody else for the initial supply. Counting that refinery as a domestic mine would make the supply picture look much more comfortable than it was. ^[2]

There are opportunities elsewhere, but the work takes time. In May 2025, Rio Tinto and Indium Corporation reported their first gallium extraction in an R&D project using material from Rio's Quebec alumina operation. The March 2026 update described a pilot plant expected to operate in 2027, with a larger demonstration plant still planned. ^[15]

A promising result must become a repeatable industrial process, then a product customers can buy in useful quantities. Quebec offers a credible route towards more supply; the announcement did not say commercial output had already arrived.

Our gallium guide follows the further journey from recovered metal to the compounds used in electronics.

Germanium and indium: follow the zinc

Germanium can come from zinc-processing residues and certain coal-derived materials. USGS has documented Chinese recovery using both domestic and imported sources. Indium is also closely tied to zinc production. The element's journey can begin in one country's geology and become a saleable product in another country's refinery. ^[5]

These aren't obscure curiosities waiting for someone to invent a use. Germanium serves fibre optics and infrared optics, among other applications. Indium is used in transparent conducting films and semiconductor products. Their customers need particular grades and forms, rather than a certificate saying a mine contains the element.

China produced about 70% of the world's refined indium in 2024. Germanium is harder to reduce to one clean percentage: current public production data are incomplete, and the USGS warns that global refinery and recycling figures are difficult to verify. We can identify China's leading role without pretending to know a precise share that the data cannot support. ^{[1, 2]}

The practical opportunity for competitors is to examine what existing zinc operations already handle. But the presence of germanium or indium in a residue doesn't make recovery profitable by itself. A recovery plant still has to deliver enough usable material to justify running.

Germanium shows what tighter exports can cost

China introduced export licensing for specified gallium and germanium products in 2023, followed by US-specific restrictions in December 2024. The rules changed access to material that factories already knew how to make. ^{[12, 25]}

USGS's European quotation for 99.999%-pure germanium metal averaged $1,392/kg in 2023, $1,991 in 2024 and an estimated $4,100 in 2025. It nearly tripled in two years. The price series does not isolate the effect of any single policy. ^[25]

The shipment figures make the squeeze more tangible. China exported 7,520 kilograms of germanium metal in January–September 2025, against 36,656 kilograms in the same nine months of 2023. That is almost four-fifths less metal reaching overseas customers. The USGS also linked a sharp fall in US germanium-metal imports in 2025 to the Chinese restrictions. ^[25]

This is the other side of the cheap-supply problem. A government can restrict particular products while other minerals remain plentiful and their prices fall. There is no single Chinese rare-metal price, and no switch that moves every material in the same direction.

Tellurium and bismuth: the small print in a big metal business

Tellurium is principally recovered from the material left during electrolytic copper refining, known as anode slimes. The name is unglamorous; the contents can be valuable. Bismuth is commonly recovered as a by-product of other metals, especially lead. ^[5]

China's estimated 2024 shares were 76% of refined tellurium and 86% of refined bismuth. Those numbers describe output from refineries. They don't mean China possesses the same percentage of every deposit containing the elements. ^[1]

Jiangxi Copper makes this connection unusually easy to see. Its company profile lists tellurium, rhenium, selenium and bismuth alongside copper, gold and silver. The group whose name brings one metal to mind participates in several smaller markets. ^[6]

That doesn't make every copper refinery a tellurium producer. Some feed is more suitable, some equipment is missing, and some recovery decisions won't make financial sense. It does explain why closing or opening a major host-metal plant can matter well beyond its headline product.

Three metals that spoil the monopoly story

Antimony is an important Chinese business, but its domestic mining share is much lower than the gallium figure: 34% in 2024, according to USGS. Chinese refineries have increasingly drawn on concentrates from Myanmar, Russia and Tajikistan. The strength of the processing business partly rests on access to somebody else's ore. ^[1]

That creates influence in both directions. Overseas customers may need Chinese refined material. Chinese plants need a dependable flow of feed. A story that gives Beijing ownership of every tonne erases the producers on the other side of those transactions.

Rhenium provides an even clearer limit. China supplied about 25% of world production in 2024, while Chile led with roughly 36%. Recovery is linked particularly to molybdenum-bearing copper ores and their processing. China's position is substantial, but it doesn't head this league table. ^{[1, 7]}

Hafnium follows zirconium. The two occur together, and hafnium is separated when zirconium is purified for applications such as nuclear fuel cladding. The United States has commercial zirconium and hafnium metal production. There is no basis here for assigning China a hafnium monopoly just because it dominates several other entries. ^[8]

These differences are useful. They show where a reader should be sceptical of sweeping claims about Chinese control. The eight-metal category helps organise the story; the individual plants determine the answer.

03 / Supplementary critical materials: graphite is a factory story too

Graphite isn't a metal. It's carbon, and it reaches battery makers by more than one route.

Natural graphite starts with mined material. To become suitable for a battery anode, it needs further processing to the required purity, shape and other specifications. Synthetic graphite is made from carbon feedstocks, including petroleum coke, through a separate industrial route involving high temperatures. Finding a graphite deposit addresses only the first of those businesses. ^[9]

China supplied about 82% of natural graphite mine production in 2024. That is already a dominant position, but it is not a measure of battery-ready anode supply. Combining natural mine tonnage, synthetic output and finished anode material in one ranking would hide the steps a battery manufacturer actually depends on. ^[1]

The supplier has to deliver material that behaves consistently inside a cell. That gives processing experience and customer qualification real value. A new mine can produce usable concentrate while its customers still rely on another country to turn it into the product their battery recipe requires.

Tungsten has more geology behind its lead

Tungsten is different. Here, China's large domestic mining industry is central: it supplied an estimated 82% of mined tungsten in 2024. The metal's hardness and heat resistance make it useful for cutting tools and other demanding applications. Further processing turns mined material into intermediate chemicals, powders and carbides. ^{[1, 2]}

So a rival tungsten project faces both a major mining incumbent and established processors. That is a different starting point from recovering gallium at an existing alumina refinery. Treating both as a generic race to open mines would send investors towards the wrong questions.

Lithium, nickel and cobalt complicate the map further. Large deposits and important processing operations sit outside China, while Chinese companies also invest abroad. A Chinese-owned interest in an Indonesian facility is different from a refinery operating inside China. A long-term purchase agreement is different again.

Those connections matter. Folding them into one unexplained percentage called “control” doesn't help anyone understand them.

04 / Base and precious metals: the businesses paying the bills

Copper, aluminium, zinc and lead get less attention in the small-metals argument than they deserve. Their much larger operations move the material from which several strategic by-products can be recovered. Gold and silver can also emerge from multi-metal processing.

The scale is easier to appreciate when mining and refining sit beside each other. In 2024, China produced about 8% of the world's mined copper but 45% of its refined copper. It mined 19% of the world's bauxite while producing 60% of its alumina. ^[1]

These are shares of separate global markets, so subtracting one bar from another doesn't calculate import dependence. The useful observation is simpler: China's processing businesses are much larger, relative to the world, than its domestic mines. Imported feed, as well as recycled material where applicable, helps support that industrial scale.

This is why “China needs imports” is not a rebuttal to its mineral strength. Buying raw material and selling a more useful product can be a strong business. It gives the processor income, experience and relationships with the manufacturers that need the output.

It does create vulnerabilities. A refinery cannot recover an ingredient missing from its feed, and it cannot run indefinitely without deliveries. Shipping disruptions, decisions by exporting governments or competition for concentrate can hurt a processor even when its machinery works perfectly.

Copper helps explain the less familiar metals

The IEA estimates that China accounted for more than 90% of the growth in global copper-smelting capacity since 2005. By 2025, it held about half that capacity. The factories expanded far beyond what the country's share of mined copper would suggest. ^[10]

Jiangxi Copper describes activities extending from mining and smelting to fabricated products and equipment. Its minor-metal output belongs within that larger business. A competitor may be able to copy a recovery process without being able to reproduce the surrounding operation cheaply. ^[6]

And those large plants must still earn money. Power, chemicals, water treatment and waste disposal remain costs. Scale can spread some costs across more output, but it doesn't abolish them. Recovering another valuable ingredient can improve the economics; it doesn't rescue every badly placed or poorly run refinery.

This matters when governments promise a complete domestic mineral industry. For some materials, the realistic alternative may be to add recovery at a plant that already exists. For others, it may require neighbouring countries to divide the work. Rebuilding every host-metal industry at home would be an exceptionally expensive way to obtain a few small outputs.

A licence can stop a factory without closing a mine

China's export controls turn industrial concentration into something buyers feel immediately. A customer doesn't need the world to run out of a metal to have a problem. It needs a delayed shipment at the moment its own stocks run low.

In 2025, that ceased to be hypothetical for some carmakers. The IEA reports that China's April rare-earth export controls forced some to reduce production or temporarily suspend operations. The ore was still in the ground and processors still existed. The material wasn't arriving where it was needed on the required schedule. ^[17]

Beijing said eligible exports could receive licences. That left the government deciding which applications qualified and buyers waiting for permission. ^[16]

A licensing system can permit trade while making it conditional. That gives a government room to approve some exports and scrutinise others. For a factory manager, the difference between a refusal and a long wait may be smaller than it looks in a press release.

The individual rules also matter. China's February 2025 order covers specified tungsten, tellurium, bismuth, molybdenum and indium items. The April order covers specified medium and heavy rare-earth items. Neither is a ban on everything containing those elements. ^[11]

The two November dates buyers shouldn't mix up

At this article's September 2026 cut-off, selected October 2025 controls are suspended through 10 November 2026. Those include the wider rare-earth measures named in Announcement 70.

A separate suspension, under Announcement 72, runs through 27 November 2026. It applies to the second paragraph of the December 2024 US-specific order, covering specified gallium, germanium, antimony and superhard-material items, plus the stricter graphite review. ^[12]

The earlier licensing systems remain relevant. The separate US military end-user and end-use restriction was not removed by that second suspension. A buyer still needs to know the actual item, destination and intended use.

That is enough legal detail for the main story, but it changes the story's meaning. “China banned rare metals” is too broad. “The restrictions were lifted” can be just as misleading. Suspensions buy time; they don't create another supplier.

What would make China's customers less dependent?

There is progress to point to. The IEA's July 2026 assessment puts the leading country's rare-earth refining share at 85% in 2025, down from over 90% in 2023, following new US projects and increased Malaysian production. The older 2024 stage comparison above remains useful because its mining, refining and magnet figures cover the same year. ^[17]

That fall is evidence that alternatives can grow. It isn't evidence that every gap has been filled. A separation plant can open while customers still lack enough metal, alloy or magnet capacity further along the route.

The same problem appears in the IEA's review of investment: projects outside dominant suppliers remain weighted towards mining, while refining and manufacturing lag behind. Governments can finance more ore and discover that they still need the same processing businesses. ^[10]

The economics of those small quantities should influence the response. A manufacturer may save a little on its material bill by buying from the cheapest established supplier. If a missing ingredient stops a much more valuable production line, a second qualified supplier can be worth paying for before the next dispute arrives.

That is a commercial choice as much as a diplomatic one. A new producer needs buyers willing to qualify the material, sign contracts and keep purchasing when the panic fades. A government can help finance a plant; it cannot make every customer treat security of supply as a permanent expense.

For by-product metals, the first useful conversation may be with the manager of an existing copper, zinc or alumina operation. What passes through the plant? Which residues contain something recoverable? What would a customer actually buy? The Quebec gallium project shows what such a route looks like, including the years between an encouraging result and industrial production.

Our rare-earth race compares the national positions. The history adds a harder test for any new supplier: can it keep operating after the shortage ends? Customers who want an alternative to China have to help make it viable during the cheap years too.