Where do rare earths come from?

In 2025 China mined 270,000 tonnes of rare-earth oxide equivalent, 69% of the world's 390,000 tonnes (USGS MCS). United States followed with 13%, then Australia (7.4%), Myanmar (5.6%) and Thailand (1.2%). The top five account for 97%. Reserves tell a less concentrated story: China holds 52% of the world's 85.3 million tonnes (USGS MCS).

Mining concentration understates the real position, and this is the single most important thing to understand about the sector. Separating the seventeen lanthanides plus scandium and yttrium from one another requires hundreds of stages of solvent extraction, because the elements differ so little chemically. That capacity is overwhelmingly Chinese, and so is the magnet-making capacity downstream. A country can mine rare earths without being able to do anything useful with them, which is exactly the position most non-Chinese producers have been in.

The deposits themselves come in a few types: carbonatites such as Bayan Obo in Inner Mongolia and Mountain Pass in California, which are rich in the light elements cerium, lanthanum, neodymium and praseodymium; ion-adsorption clays in southern China and Myanmar, which are the main source of the heavy elements dysprosium and terbium that magnets need for heat resistance; and mineral sands containing monazite, a by-product of titanium and zirconium mining.

Production means mined oxide equivalent, not separated oxides and certainly not magnets. World output changed +3% on the previous year, and the trend has been upward as non-Chinese mines opened, but the separation bottleneck has moved much less.

China was the largest exporter of rare-earth metals and compounds (HS 2805) in 2024 with 38% of world export value, ahead of Vietnam (18%), on world trade of $1.1 billion (CEPII BACI). Japan was the largest importer with 25%.

Trade statistics for rare earths are genuinely difficult to read. The customs headings mix unseparated concentrates, individual separated oxides worth vastly different amounts, metals and alloys, so a value share does not translate into a quantity share. A large share of world trade is also intra-industry: concentrate moves to China for separation and comes back as oxide or metal, and finished magnets move again under an entirely different heading. The map that matters most, magnet production, does not appear in this table at all.

This page does not quote a rare-earth price, and the reason is worth stating plainly. There is no exchange benchmark and no free, citable public price series. Individual separated oxides, neodymium-praseodymium, dysprosium, terbium, lanthanum, cerium and the rest, are assessed by private price reporting agencies whose data is licensed, and the Chinese domestic market and the export market can diverge substantially. Publishing a number here would mean either paying for a license this site does not have or making one up, so it does neither.

What can be said is how prices are actually set. Most material moves under annual or multi-year contracts between separators and magnet makers, referenced to a private index. Chinese production and export quotas, set administratively, have historically been the largest single influence on the level. Because the elements come out of the ground in fixed proportions, producers face a balance problem: demand is concentrated in neodymium, praseodymium, dysprosium and terbium for magnets, while cerium and lanthanum come out in far greater quantity than anyone wants and are effectively waste with a disposal cost. The economics of a mine therefore depend on its basket composition, not on a single price.

For readers wanting a directional sense of the market, the volumes and shares on this page, the concentration of separation capacity, and the policy timeline below say more about cost and availability than any single quotation would.

Neodymium-iron-boron magnets in electric vehicle motors, wind turbine generators, robotics, drones, and every hard drive and speaker are the demand that matters. Motors are the largest and fastest-growing use, and each direct-drive wind turbine or traction motor contains kilograms of magnet, so magnet demand tracks electrification directly.

Quotas on production and export, environmental enforcement against unlicensed mining, and export licensing on specific heavy elements have repeatedly moved the market. Because separation capacity is concentrated, an administrative decision affects world availability in a way no mine closure would.

The elements are mined together in fixed geological ratios, but demand is wildly uneven. Producing enough neodymium means producing far more cerium and lanthanum than the market wants, so surplus light elements are stockpiled or discarded. A mine's viability depends on whether its ore is rich in the elements people actually pay for.

Dysprosium and terbium, which let magnets hold their strength at motor operating temperatures, come mainly from ion-adsorption clays in southern China and Myanmar. That supply is small, geographically narrow and exposed to conflict and enforcement, and it is the tightest part of the chain.

Building a separation plant is a chemical engineering project of hundreds of stages, with difficult effluent and, where monazite is involved, low-level radioactive thorium residues to manage. Permitting is the binding constraint, which is why announced non-Chinese capacity has taken far longer to arrive than announced non-Chinese mines.

Motor designers reduce dysprosium content through grain-boundary diffusion, redesign magnets to use less material, or adopt ferrite magnets and externally excited rotors that avoid rare earths entirely. Each price spike accelerates this work, and the demand lost does not fully return when prices fall.

Magnets from end-of-life vehicles, wind turbines and hard drives contain concentrated, already-separated material, and recovering them avoids the separation step entirely. Volumes remain small because collection and disassembly are difficult, but this is the most plausible route to non-Chinese supply of heavy elements.

Mining is conventional: open pit at Bayan Obo and Mountain Pass, in-situ or heap leaching of ion-adsorption clays, or recovery of monazite from mineral sand dredging. The ore is concentrated by flotation or gravity to a mixed rare-earth concentrate, typically cracked with acid or alkali to break down the mineral structure and put the elements into solution.

Then comes the step that defines the industry. Because the lanthanides differ by a single electron in an inner shell, their chemistry is almost identical, and separating them requires solvent extraction in cascades of hundreds of mixer-settler stages, each achieving a tiny enrichment, repeated until the individual oxides reach the required purity. The capital is moderate but the process knowledge, the effluent handling and the permitting are not, and this is where the world's dependence actually sits.

Separated oxides are then reduced to metal, usually by molten salt electrolysis or metallothermic reduction, and alloyed with iron and boron. The alloy is jet-milled to powder, aligned in a magnetic field, pressed, sintered, machined and coated to make a finished magnet. Heavy elements may be added at the grain boundary rather than throughout, which achieves the same heat resistance with far less dysprosium.

Monazite-bearing ores carry thorium, which is mildly radioactive, so residues require licensed disposal. This is a real and often decisive obstacle to permitting separation plants outside China, and it explains part of why the industry consolidated where it did.

Permanent magnets are the dominant use by value: neodymium, praseodymium, dysprosium and terbium go into the magnets that drive electric vehicle traction motors, direct-drive wind turbine generators, industrial robots, actuators, hard drives, headphones and countless small motors. This is the use that makes rare earths strategic.

Beyond magnets, cerium is used as a glass polishing compound and in catalytic converters, lanthanum in petroleum fluid cracking catalysts and in nickel-metal-hydride batteries, europium and terbium as phosphors in lighting and displays, yttrium in high-temperature ceramics and superalloys, gadolinium in medical imaging contrast agents and neutron shielding, and samarium in high-temperature magnets for aerospace and defense. Erbium is essential to the optical amplifiers that make long-haul fiber networks work.

The defense applications, in precision guidance, radar, sonar and aircraft actuators, are small in volume and disproportionate in strategic weight, and they are the reason governments treat rare earth supply as a security question rather than a commercial one.

The chain runs mine, concentrate, separation, metal, alloy, magnet, and China holds a commanding position at every stage after the first. That is the whole story of rare earth supply risk: opening a mine elsewhere does not reduce dependence if the concentrate must still be shipped for separation, and building a separation plant does not help if there is no magnet maker to buy the oxide.

The tightest single point is heavy rare earth supply from ion-adsorption clays. Those deposits are geographically small, environmentally damaging when mined informally, and partly located in areas affected by conflict in northern Myanmar, from which material crosses into China for processing. A disruption there affects the elements with the fewest substitutes.

Thorium disposal is the quiet constraint on diversification. Any separation project handling monazite must find a licensed route for low-level radioactive residue, and in most jurisdictions that is a multi-year permitting exercise with strong local opposition. It is the reason several announced non-Chinese projects have slipped repeatedly.

The market's own economics are a risk too. Because prices are volatile and administratively influenced, and because a new entrant must sell a full basket including elements nobody wants, financing a mine-to-magnet chain outside China has usually required a government to underwrite part of it. Where that support has been available, projects have advanced; where it has not, they have stalled regardless of the geology.