Are rare earths rare? Why 17 elements take 1,000 steps to separate

There are 17 rare earth elements, and the common ones sit in the crust at about the same concentration as copper.² There is more cerium in the Earth's crust than copper, and roughly as much neodymium as copper. Even thulium, the scarcest stable rare earth, is nearly 200 times more common than gold.³

The US Geological Survey calls the name a "historical misnomer" in its own fact sheet.⁴ The name comes from 18th century chemistry. "Earth" was the old word for a metal oxide. "Rare" came from a Swedish quarry, where the first of these earths was identified in 1794 in a mineral that looked scarce.⁵

The world is not running out

Everything ever mined in history adds up to about 5.9 million metric tons of rare earth oxides. Stated reserves today are more than 75 million tons, which is about 12.6 times everything we have ever pulled out of the ground.⁶

So the problem was never finding them. The problem is that geology hands them to you glued together.

The reason the elements come glued together is a quirk in where the electrons sit. As you move along the row of the periodic table, each new electron drops into an inner shell underneath the outside of the atom. Chemistry only ever sees the outside, so from the outside all 17 look almost the same.⁷

One thing does change. Each element is very slightly smaller than the one before it. Neodymium and praseodymium, the pair the magnet industry cares about most, differ in size by 1.8%.⁸

That 1.8% is the whole lever. Every separation plant on Earth exists to turn it into a 99.9% difference in purity.⁹

The method is called solvent extraction. You wash the dissolved ore against an oily liquid that very slightly prefers the smaller ion. One pass leaves you about 1.5 times purer than you started.¹⁰ In the chemist's terms, the oily side ends up about 1.5 times richer in neodymium, relative to praseodymium, than the watery side.

Then you do it again, and again. Chinese researchers at the General Research Institute for Nonferrous Metals put the real number in a 2015 paper: more than 1,000 stages of continuous extraction to separate the individual elements.¹¹

That count is the scarce part of rare earths. Every one of those stages has to hand the next a slightly purer liquid, and the whole chain rests on a size difference of 1.8%.

The rock still travels to the chemistry

Until the middle of 2025, America's only rare earth mine shipped most of what it dug to China for processing, because that is where the separation plants are. MP Materials says so in its own filing, and it stopped those sales in July 2025.¹²

MP Materials' annual filing for its Mountain Pass mine in California puts a price on the two halves of the business. Mining a short ton of ore costs $1.50. Concentrating that same ton of ore costs $51.28. Separating a short ton of the concentrate that comes out costs $1,080.59.¹³

The digging is a rounding error against the chemistry.

A short ton is 2,000 pounds, the unit MP Materials reports in. The three numbers sit in the same table of the same filing, so the comparison is the company's own. The first two lines are the mine. The third is the chemistry, and it is the largest of the three by far.

This is the lens I keep through the whole rare earth series. In the video I close on it: the mine was never the moat. The chemistry was.

Inside the 17 there is one split that matters more than the rest, light against heavy. The light ones, lanthanum, cerium and neodymium, are the common ones, and the world has found plenty of them outside China. The heavy ones, dysprosium and terbium, are the ones a magnet needs to survive heat, and they come overwhelmingly from clay deposits in southern China and Myanmar.¹⁴

The price shows the split. In 2025 lanthanum oxide sold for about $1 a kilogram. Terbium oxide sold for $1,010.¹⁵ Both come out of the same rock, one element at a time.

The heavy elements are a small share of every rock

Opening more mines helps less on the heavy side, because the heavy elements are a small share of the rock. At Mount Weld in Western Australia, the deposit Lynas mines, lanthanum and cerium, the two cheapest elements on the board, are 68.7% of the reserves. Dysprosium and terbium together are 0.78%.¹⁶ It is the same rock from the same hole. The difference is how much of each element the rock holds.

Mountain Pass shows even less on paper. Its resource statement leaves the heavy rare earths out, and the filing says they "are not accounted for in this estimate due to historic data limitations."¹⁷

Heat decides which magnets need the heavy elements. A speaker is a static case: the magnet sits still. In a dynamic case such as an electric motor, heat and the magnetic field become a concrete problem, and that is where dysprosium and terbium come in. I cover them in dysprosium and terbium, the heavy rare earths that decide electric motors.

China has drawn the same line. In April 2025 it put 7 rare earth elements under export license, and apart from samarium they sit at the heavy end of the table.¹⁸ The dates that followed, and the ones still to come, are in China's rare earth export controls timeline.

What to do with this: the next time you read about a new rare earth mine, ask two questions before you read on. Who separates the ore, and does it carry the heavy elements? It is the question I ask of every layer in the series: where are the bottlenecks, and who sits on them? For rare earths the bottleneck is the separation plant, and the companies that run one outside China, or make magnets there, are in rare earth companies outside China.