Chinese separation, Ganzhou & Baotou
The bottleneck is not the mine but separation: solvent-extraction circuits with hundreds of stages, optimised over forty years. For dysprosium and terbium, China holds an almost absolute monopoly. No Western substitute reaches comparable industrial scale before the end of the decade.
Separation/refining remains the true chokepoint: China supplied 91% of global refined rare-earth output in 2024, and the IEA projects that non-China capacity will meet only about 25% of refining demand by 2035, leaving dependence largely intact.AIE, Rare Earth Elements, 2025 ↗
China's dominance over rare earths is decided at separation, not at the mine. Chemically separating near-identical lanthanides from one another requires hundreds of stages of counter-current cascades, a know-how China spent half a century perfecting and that no one has replicated at scale. Diversifying mines is not enough: without separation capacity outside China, any new ore flows back to feed the monopoly.
The dependence is not created at the mine (60 %) but at refining (91 %); it peaks on heavy rare earths (99 %) and magnets (94 %). The lock is chemical and industrial, not geological. Sources: IEA 2024, USGS 2025, CSIS. Metallization (88 %) is a sector estimate.
The power lies in the chemistry, not the mine
Rare earths abound in the Earth's crust; what is missing is the ability to separate them. Chemically near-twins, the lanthanides differ only at the margin: a single stage of solvent extraction barely separates them at all. A plant producing several pure elements at more than 99.99 % therefore lines up hundreds of mixer-settler stages in counter-current cascades. Market power is concentrated there, not at the mine.
China extracts about 60 % of the world's ore, but processes 91 % of it, and up to 99 % of the heavy rare earths, those that hold heat in an electric motor or a missile actuator. The gap is technical, not geological. The process itself rests on the theory of counter-current cascade extraction formulated in the 1970s by Xu Guangxian, the "father of Chinese rare-earth chemistry," who collapsed the cost and time of separation and founded the country's dominance.
Why the monopoly holds
Forty years of cumulative optimization, first: as early as 1995, China already produced 90 % of high-purity rare-earth metals. Half a century of industrial learning that no competitor has caught up on, the West closed its lines, let its engineers leave and lost the process R&D.
The environmental cost, next, was externalized. Separating one ton of rare earths generates on the order of one ton of radioactive thorium residues, thousands of cubic meters of gas and acidic wastewater. The Baotou tailings pond, built without an impermeable membrane, is contaminated with lead, cadmium and thorium. The West refused this liability; China absorbed it, hence part of its cost advantage.
The price weapon, finally. Historically, whenever a Western project emerged, China flooded the market to crash prices and destroy the investment case. Even adding up all announced projects, the International Energy Agency projects that refining outside China will cover only a quarter of 2035 demand.
The Achilles' heel: the South and Myanmar
Chinese dominance itself has a weak point. Heavy rare earths do not come from the great Bayan Obo deposit (in the North, near Baotou), which is essentially light, but from the ionic clays of the South, Jiangxi, around Ganzhou, and, increasingly, from neighboring Myanmar. In 2023, Myanmar supplied about 98 % of China's imports of heavy oxides: Ganzhou's separation depends on ore extracted in a conflict zone of Kachin State, beyond any stable control.
In other words, the most critical node of the energy transition rests on a chain of two fragile links: a know-how concentrated in Ganzhou, and a raw material partly imported from a country in civil war. A shock to either propagates instantly to the world's magnets.
The redundancy being built, slowly
Outside China, separation accounts for about 10 % of the total, and near zero for the heavies. Three players are emerging. Lynas (Australia/Malaysia) is the leading producer outside China and, since May 2025, the only commercial producer of separated heavy rare earths outside China, but at small scale and under financial strain. Solvay inaugurated a magnet line in La Rochelle in April 2025, still marginal. MP Materials (California) shipped, until 2022, most of its concentrate… to China, to be separated there.
States are remedying this with public capital: the Pentagon became MP Materials' largest shareholder in July 2025, with a guaranteed floor price over ten years; Australia backs Iluka with a 1.25 billion dollar public loan. But building and qualifying separation capacity takes years, and the hardest link, the heavies, remains almost entirely Chinese. Redundancy will exist, in the minority, on the scale of the decade.