Two Prices for the Same Kilogram

Imagine you run a magnet factory in Europe, and you need dysprosium oxide — the heavy rare earth that keeps a magnet working at the temperatures inside an electric motor. In the summer of 2026, the price quoted to you in Rotterdam was $1,250 to $1,675 a kilogram. The same material, quoted free-on-board in China, was $270 to $330 a kilogram. Same oxide. Same purity, roughly. Five times the price, depending on which side of a border you stand on. The market for rare earths is not one market. It is two markets, separated by a licensing desk in Beijing.

The split began on April 4, 2025, when China's Ministry of Commerce issued Announcement No. 18, imposing export licences on seven rare-earth elements — samarium, gadolinium, terbium, dysprosium, lutetium, scandium, and yttrium — and the magnets made from them. It was not a ban. It was a licence regime, and a licence regime is worse than a ban in one respect: nobody knows how long a licence takes. By May 2025, reported Chinese exports of dysprosium oxide had fallen to zero, and magnet exports were down 76% year over year. The taps had not been closed. They had been put on a timer nobody could read.

01 — The gap widens, element by element

The October 2025 package made the pattern explicit. On October 9, China added five more elements — europium, holmium, erbium, thulium, and ytterbium — plus controls on rare-earth equipment and technology, and a rule that any product containing more than 0.1% Chinese-origin rare earths would itself need a licence. The terbium gap was already stark: $1,180 to $1,270 a kilogram free-on-board in China against $4,200 to $5,000 in Rotterdam, roughly four times. For samarium, the indicative gap ran wider still — Chinese quotes near $11 to $13 a kilogram against $3,000 to $4,800 outside it — though that figure rests on thin, lightly traded quotes and should be read as a signal of dislocation, not as a clearing price.

These are not prices set by supply and demand in the usual sense. They are prices set by access. China mined about 60% of the world's magnet rare earths in 2024, refined about 91% of them, and made about 94% of the world's sintered neodymium-iron-boron magnets. When a single country refines nine-tenths of a material, the price inside that country and the price outside it are two different facts. The International Energy Agency warned on November 1, 2026 that prolonged restrictions could threaten $6.5 trillion in annual manufacturing production outside China — a figure that sounds abstract until you remember that it includes the motors, the turbines, and the defense systems that run on these elements.

02 — Truces, suspensions, and the clock

The controls have not stood still. The October 2025 package was suspended on November 7, 2025, as part of a U.S.–China trade truce, with the suspension running to November 10, 2026. In September 2026, Treasury Secretary Scott Bessent said the suspension would be extended into January 2027, and China's Ministry of Commerce confirmed that the wider U.S. trade arrangements would run to January 10, 2027 — though that statement did not name rare earths, and no numbered notice changing the November 10 date had been published as of mid-October. The April 2025 controls on the original seven elements, notably, were never suspended at all. The licensing regime for dysprosium and terbium has now been in force, continuously, for more than eighteen months.

The calendar matters because the response is measured in years, not quarters. MP Materials expects its NdPr production run rate to exceed 5,000 tonnes a year across 2027, and its heavy rare-earth separation — the part of the chain that matters most for dysprosium and terbium — is only beginning commercial production in the fourth quarter of 2026. Lynas Rare Earths produced 7,260 tonnes of NdPr in its 2026 financial year, and first produced samarium oxide in March 2026. These are real milestones, and they are also a measure of how long the gap will persist: even the fastest builders in the West are years from closing it.

That is the arithmetic that makes the ex-China premium a business case rather than a grievance. A mine or a separation plant outside China does not have to beat China's cost. It has to beat the Rotterdam price — the price a buyer actually pays when the licence queue is slow. At $1,250 to $1,675 a kilogram for dysprosium oxide outside China, against $270 to $330 inside it, a Western producer has room to be far less efficient than its Chinese counterpart and still be the cheaper supplier to a European or American buyer. The premium is the margin. And a royalty on that producer's output is a claim on the premium itself.

03 — The honest caveat

There are good reasons not to take the Rotterdam quotes at face value. They are indicative prices from thin markets — a handful of trades, a handful of quotes — and a thin market can show a five-times gap on a Tuesday and a two-times gap on a Friday. The controls, too, are licences rather than embargoes: material does flow, with paperwork, and the flow has improved since the worst months of 2025. The truces have held, so far. And the premium creates its own undoing: the higher the ex-China price, the more capital rushes into ex-China supply, and the more supply there is, the narrower the gap becomes.

But the direction of travel is set by the structure, not by the headlines. China still refines about 91% of magnet rare earths. The United States still mines 13% of the world's rare-earth oxides and refines a fraction of that. The IEA still expects total rare-earth demand to rise from 91 kilotonnes in 2024 to 123 kilotonnes by 2030, with clean-energy demand roughly doubling. And 20.7 million electric cars were sold in 2025, each one carrying kilograms of magnets that cannot be made without these elements. As long as one country refines nine-tenths of the supply and the rest of the world buys a quarter of its cars electric, there will be two prices for the same kilogram — and the gap between them will keep financing the mines, the separation plants, and the royalties that close it.