The History of Rare Earths

The numbers tossed around when talking about rare earths are actually not just of one kind. At least four completely different figures circulate under the very same name: the amount buried in the ground, the amount mined from the ground, the amount separated into the seventeen individual elements, and the amount turned into magnets.

According to a table published by the U.S. Geological Survey in February 2026, global mine production in 2025 was 390,000 tons in rare earth oxide equivalent, with China accounting for 270,000 tons. Do the math, and that comes out to 69%. When the International Energy Agency counted only the elements used in magnets, China's mining share in 2024 was about 60%. Yet according to that same agency, China's share at the separation and refining stage is 91%, and for sintered permanent magnets, it is 94%.

Such a wide gap between these numbers tells us that the real center of gravity in this story is not "where it is buried," but "who separates it." In fact, over the past decade or two, the number that dropped significantly was the mining share; the number that remained virtually unchanged was the separation share.

That is why, throughout this series, we will specify which stage we are talking about every time a number comes up. A statement like "China controls 90%" is only half true unless you specify which step of the process you mean.

Let us clarify one detail beforehand. The "tons" used in these tables usually refer to weight in rare earth oxide equivalent. The figures shift if measured in pure metal, and they vary depending on whether you count all seventeen elements or only the handful used in magnets. That is why we will specify which agency counted each number, when, and by what standard. Without those three things—the agency, the point in time, and the scope—percentages cannot truly be compared with one another.

News reports often show maps partly because maps are easy to draw. Factories, on the other hand, rarely show up on a map. Yet in this industry, what truly decided the game was the part that never appears on a map.

This statement comes up often, and it is half true. You have to break down which axis it holds true on to see the other half.

The first axis is how abundant they are in Earth's crust. On this front, it is true. According to tables from the U.S. Geological Survey, cerium exists in the crust at 66.5 ppm, neodymium at 41.5 ppm, and lanthanum at 39 ppm. Cerium is more common than copper or lead. Even thulium and lutetium, the least abundant among them, are about 200 times more common than gold. In a 2002 publication, the same agency stated definitively that the label "rare" is "a historical misnomer."

The second axis is whether they are concentrated enough to be worth mining. Here, it is false. The 2026 edition from the same agency notes that "rare earths are relatively abundant in the Earth's crust, but minable concentrations are less common than for most other mineral commodities." Being widely dispersed and gathering together in one place are two completely different things.

The third axis is the central axis of this story. What comes out of the ground is a lump where seventeen elements are jumbled together. What we can actually use are the elements separated one by one. The lumps are common; what is rare are the facilities that can separate that lump.

Calling the name itself a "historical misnomer," the U.S. Geological Survey notes that the term lingers on not because these elements are genuinely rare, but because they are unfamiliar. In the same paragraph, it also compares the more abundant rare earths, noting that their crustal concentrations are similar to common industrial metals like chromium, nickel, copper, zinc, molybdenum, tin, tungsten, and lead.

There is also more than one way to count them. Usually, people count the fifteen lanthanides plus yttrium, and sometimes add scandium to make it seventeen. Footnotes in the U.S. Geological Survey's rare earth statistics clarify that they include the lanthanides and yttrium, while mostly excluding scandium. Each element's price and utility also vary dramatically: lanthanum through gadolinium are classified as light rare earths, terbium through lutetium as heavy rare earths, and yttrium is grouped with the heavy rare earths because of its similar properties.

The reason rare earths are so hard to separate is that they are chemically too similar to each other. When dissolved in water, neighboring elements behave in almost the exact same way.

Praseodymium means "green twin," and neodymium means "new twin." For these two, which are twins even by name, the separation factor is not even 2 with most extractants. That means when you mix them and let them settle once, one side becomes only infinitesimally more concentrated.

So the method used is repeating the same step over a long sequence. Tanks that mix the aqueous solution with an oil-phase solvent and let them settle into two layers are called mixer-settlers. You connect these in a long series and flow the liquids in opposite directions. A tiny bit separates in one tank, and a little more in the next. An academic review published in 2014 notes that they are sometimes linked across hundreds of stages to achieve the required separation. Because this value varies greatly depending on the ore and the target element, there is no single number for "how many stages is typical."

Even from this much, you can tell: this process is not the kind of thing solved by a single invention. It works only when you have a factory with long rows of tanks, alongside people who know how many tanks to set up and how much liquid to flow through them.

This is also why heavy rare earths are so much more precious. Dysprosium and terbium, which are used for high-temperature performance in magnets, are present in far smaller amounts than light rare earths, and separating them is even more difficult. In 2025 annual average figures, dysprosium oxide was $239 per kilogram, and terbium oxide was $1,010. In that same year, cerium oxide was $1.71, and lanthanum oxide was $1.00.

So it is easy to see that moving this process to another country is entirely different from simply buying machines. Which extractant to use on which ore, across how many stages, and how to flow the liquids varies by ore body—and the answers only accumulate with those who have actually run the operation.

Let's start the story in the gas lamp era. Before electricity arrived, the item that dramatically boosted the brightness of gas lamps was the incandescent gas mantle. When a mesh-like fabric treated with metal oxides was placed over a flame, it glowed bright white.

The composition of that oxide is where our story begins. A paper published in 2015 describes it as 99% thorium oxide and 1% cerium oxide. What was actually being sold was thorium. To obtain thorium, processors had to refine a mineral called monazite, which was packed with rare earth elements. But back then, there was virtually no market for rare earths, so in the words of that same paper, they were treated "almost as a waste product."

Today, the exact opposite is true. Rare earths are the product, while thorium is hazardous waste. Thorium itself is only weakly radioactive, but daughter radionuclides like radium can accumulate during processing. The U.S. Geological Survey notes that this concern essentially pushed monazite out of its spot as a primary raw material. As a result, attention shifted to bastnasite, a mineral low in thorium.

Within a single century, what counted as the product and what counted as the burden in that very same mineral completely flipped. This reversal sets the stage for every chapter that follows. After all, deciding who handles this process ultimately comes down to who is willing to take on what is left behind.

That is why international trade in that era revolved around monazite. The Congressional Research Service summarizes that although monazite was the most common rare earth mineral, because it contains thorium, high processing and disposal costs effectively removed it from the raw material lineup in the United States. In its place, the low-thorium mineral bastnasite stepped in.

In short, it comes down to this: in this industry, who builds a plant and where has always been decided not so much by where the ore is located, but by who can stomach what comes along with it.

In 1949, an ore body remarkably rich in light rare earths was discovered at Mountain Pass in the Mojave Desert of California, United States. Its rare earth oxide content was 8~12%. Yet a discovery alone does not keep factories running. They needed buyers.

That customer was color television. Europium was used for the red phosphors on screens, and there was no substitute. The U.S. Geological Survey notes that what supported the initial development of Mountain Pass was europium demand created by the commercialization of color television. The Mountain Pass era began in 1965, and by around 1966, this single ore body had become the world's largest supplier.

Yet right here, a problem unique to this industry reared its head. You cannot simply cherry-pick and mine europium alone from the ore. The europium content of Mountain Pass bastnäsite is about 0.1 weight percent. According to calculations in a 2015 paper, to obtain 1 ton of europium oxide from that ore, 450 tons of cerium oxide, 300 tons of lanthanum oxide, and 118 tons of neodymium oxide come out with it.

This is called the "balance problem." To get one element, you end up with mountains of others that have nowhere to be sold. So uses had to be found for the leftover elements as well. Cerium was in a better position among them. The U.S. Geological Survey notes that cerium oxide is uniquely suited as a glass polishing agent, and polished glass products—from everyday mirrors and eyeglasses to precision lenses—are finished with cerium oxide. Conversely, elements that occur in tiny quantities in the ore see their prices soar dramatically with even a slight shortage.

One more thing. In bastnäsite ore, all elements from europium through lutetium combined with yttrium make up only 0.4%. The dysprosium and terbium needed for modern magnets are tucked inside that very 0.4%. That means it is difficult to obtain enough heavy rare earths from this ore, which is why the ion-adsorption clays of southern China will become so special in the story ahead.

In Chinese records, there is a scene that clearly illustrates the nature of this industry. In September 1974, an industrial trial separating elements like praseodymium and neodymium was carried out by connecting eighty stages of mixer-settlers. The result was neodymium, praseodymium, and lanthanum with purities in the 99% range.

What came next was even more important. In the late 1970s, Xu Guangxian of Peking University formulated the "Cascade Countercurrent Extraction Theory." Until then, the required number of stages and liquid flow ratios had to be found through trial and error, shaking separatory funnels by hand. This theory made it possible to derive those values through calculation. Process design shifted from fumbling by hand to solving on paper.

This kind of change is hardly visible on the surface. After all, no new material was invented, nor was a new mine opened. Yet in this industry, breakthroughs like this are decisive. As we saw in the previous chapter, the bottleneck was never the basic principle itself, but how many stages to connect and how to arrange them.

According to the timeline of the U.S. Geological Survey, the years from around 1984 to 1991 mark a transition period, and after 1985, China's production grew sharply. However, the claim that "outside factories shut down because of this theory" comes from a retrospective in a Chinese academic journal. It is worth keeping in mind that this is an account written by one of the involved parties.

The U.S. Geological Survey divides global rare earth production into four eras. The monazite placer era ended in 1964; the Mountain Pass era ran from 1965 to the mid-1980s; followed by a transition period, leading into the China era from 1991 onward. What this division tells us is that dominance did not shift by following the raw resources—it shifted by following the processing methods.

The retrospective in the Chinese journal notes that after 1995, China came to account for more than 90% of global production and trade in high-purity, individual rare earth products. It is worth noting that what is being referred to here is not raw ore, but "individually separated products."

This is the part of the story that is most often misunderstood. It is commonly summarized as "cheap Chinese imports forced the American mine to close." But the actual sequence was different.

Here is how the U.S. Geological Survey described it in 2002: following environmental and regulatory issues surrounding a wastewater pipeline, the solvent-extraction separation plant was shut down, after which the mine produced only bastnäsite concentrate, selling separated products out of inventories built up before the shutdown. The separation plant stopped in 1998, and the mine stopped in 2002. There was a 4-year gap.

What happened in the meantime is also recorded in that same document. In 1999 and 2000, more than 90% of the rare earths consumed by U.S. industry came from China. That happened even while the mine was still running. Even if you retain the ability to mine, if you lack the ability to separate, you become an importer.

An April 2010 report from the U.S. Government Accountability Office, cited by the Congressional Research Service, illustrates this structure step by step. At that time, China accounted for about 95% of raw materials, about 97% of oxides, about 90% of metal alloys, 75% of neodymium magnets, and 60% of samarium-cobalt magnets. The pattern where China's share grew as you moved down the supply chain was already distinct back then.

The exact same pattern persists today. In a February 2026 table from the U.S. Geological Survey, the United States mined 51,000 tons of concentrate in 2025, but produced 8,900 tons of compounds and metals. The figures for extracting and separating remain on completely different orders of magnitude.

So in this industry, "securing resources" and "securing supply" do not mean the same thing. The former is mining rights; the latter is a plant lined with tanks.

And this distinction is the key to understanding what is unfolding today. Export controls, price floor contracts, and permitting requirements all target that processing plant next door far more heavily than the mine itself.

Why, then, did other countries let go of this process? Instead of pointing to a single cause, the U.S. Geological Survey lists four factors side by side: China's lower labor and regulatory costs, the continuous expansion of Asian manufacturing, the number, scale, and heavy rare earth content of Chinese deposits, and environmental and regulatory issues at Mountain Pass. The same document also points to the transfer of processing and application technologies from the United States and Europe to Asia.

The same dynamics were at work on the raw materials side. The U.S. Congressional Research Service notes that radiation concerns over thorium in monazite effectively pushed this mineral out of its role as a primary rare earth feedstock in the United States. That was because the cost of treating and disposing of thorium was so high. There are also accounts stating that overseas separation plants were forced to close, but it is worth noting that this perspective comes from Chinese historical recollections.

Meanwhile, Chinese recollections explain the shift through prices, while the U.S. Geological Survey points to wastewater pipelines and regulatory hurdles. Both accounts describe the exact same period. If you ignore one side, you are left with only half the picture.

It is also worth noting that these calculations are not just relics of the past. In March 2026, Malaysia extended the operating license of a rare earth processing facility in the country by 10 years, but attached strict conditions: Stop producing radioactive residue within 5 years, and neutralize residues generated before then through methods such as extracting thorium.

The conscious choice to avoid thorium even reshaped the types of raw materials used. As a result, the low-thorium mineral bastnäsite became the primary feedstock. Countries did not step away simply because prices were low, nor because they lacked the technology. It was simply that, on their balance sheets at the time, this process was no longer a profitable business.

In this chapter, we won't judge "which country made the right move." Each country simply had its own items on its own ledger. However, the fact that there were items left off those ledgers will become clear in the next chapter through the official documents of the nations involved.

On this point, it is more accurate to read a document published by the Chinese government itself rather than relying on outside assessments. In June 2012, the State Council Information Office released a rare earth white paper.

The document describes its own role like this: supplying more than 90% of global demand with about 23% of the world's reserves. It also notes that the smelted and separated products made in 2011 totaled 96,900 tons, accounting for more than 90% of world production. Here again, we see that the numbers on the reserve axis and the separation axis are vastly different.

The latter part of the same document reads like an itemized bill. It records that producing 1 ton of rare earth oxide using older leaching methods in southern ion-absorption mines generated 2,000 tons of tailings. Resource recovery rates were less than half, and in Baotou, only one-tenth of reserves were beneficiated and utilized. The document also diagnosed massive overcapacity in smelting and separation, listing landslides, clogged rivers, and major pollution incidents.

There are also passages addressing price. For years, the price of rare earths failed to reflect resource scarcity and environmental damage. The paper set out a stark comparison: while rare earth prices rose 2.5 times from 2000 to 2010, gold rose 4.4 times, copper 4.1 times, and iron ore 4.8 times. When environmental costs are not built into the price, those costs do not vanish—they simply settle somewhere else.

The future direction outlined in this document is equally worth noting. The central government would set and allocate aggregate quotas for mining and production, consolidate enterprises, and establish new emission standards. It was an intention to rein in market overcapacity through state administrative power. Here, we see a key difference from other heavy industries. In steel, excess capacity lingers because shutting down equipment is so difficult; here, instead of turning everything off, authorities set overall quotas to choke down the flow.

In 2010, China began tightening rare earth supplies through export quotas, export licenses, and taxes. Rather than using tariffs, it controlled the flow through physical volumes and permits.

Prices swung dramatically. According to figures from Japan's Ministry of Economy, Trade and Industry cited by the U.S. Congressional Research Service, dysprosium metal surged from $250 per kilogram in April 2010 to $2,840 in July 2011. Neodymium metal rose from $42 to $334 over the same period.

The regulatory wheels began to turn as well. On March 13, 2012, the United States brought a dispute against these measures before the World Trade Organization, joined by Japan and the European Union filing complaints on the same matter. This was followed by a panel report on March 26, 2014, an Appellate Body report on August 7, and its formal adoption on August 29. China subsequently notified the WTO on May 20, 2015, that it had eliminated export duties, export quotas, and restrictions on the right to export.

Yet prices dropped well before the rules did. The U.S. Congressional Research Service noted that prices of rare earth oxides and metals fell in the first half of 2012 and declined further in the second quarter of 2013. The primary reason cited was not new supply, but the demand side: partial substitution, accumulated stockpiles, and a sluggish economic recovery. It pointed to both China's export limits and the absence of production capacity elsewhere as the forces that had driven prices up in the first place. The rules did not actually change until roughly two years later.

Here, a key difference from the steel industry emerges. In steel, blast furnaces cannot simply be shut down, so overcapacity lingers; rare earth separation facilities, on the other hand, were shut down multiple times. Instead, it takes time to turn them back on. Mountain Pass was transferred to new owners in 2017 and resumed concentrate production the following year, but producing separated products took several more years. In the meantime, overcapacity inside China was held in check not by market forces, but by total-quota controls and corporate consolidations.

This was when many countries realized for the first time that supply could be dictated by rules rather than market prices. And that was precisely what the World Trade Organization addressed—not tariffs, but that very mechanism.

Let’s return to the four initial pillars. In a table published by the U.S. Geological Survey in February 2026, global mine production in 2025 reached 390,000 tons, with China at 270,000 tons, the United States at 51,000 tons, Australia at 29,000 tons, and Myanmar at 22,000 tons. However, a footnote in the same table points out that figures for countries such as Myanmar, Thailand, and Madagascar are not official reports from those nations, but values estimated backward from China's import statistics.

Reserves are the softest figures in this table. The global total in the 2026 edition stands at over 75 million tons, compared to over 90 million tons in the previous edition. This change did not happen because ore underground suddenly vanished, but because estimates were reassessed to reflect new reporting from national governments and mining companies. The figures for Brazil, Vietnam, India, and Malaysia were revised entirely.

Conversely, we are seeing slight shifts in an axis that had long remained stagnant. Rare earth compounds and metals produced by the United States rose from 800 tons in 2023 to 4,300 tons in 2024, and 8,900 tons in 2025. In its 2026 outlook report, the International Energy Agency highlighted rare earth refining as an exception where market concentration had actually edged down slightly, pointing to new U.S. facilities and expanded output in Malaysia. Downstream at the magnet stage, however, the trend moved the other way. The same agency notes that China’s share of sintered permanent magnets reached 94% in 2024, up from around half roughly two decades ago. Among the four axes, this one did not decline—it climbed.

Policy frameworks are shifting as well. On October 1, 2024, China implemented the Rare Earth Administration Regulations, bringing mining, smelting and separation, distribution, and imports and exports under a single legal umbrella. In April 2025, it introduced export controls on seven elements and magnets, expanded the scope in October, and then granted a one-year suspension on the October expansion in November. Meanwhile, the European Union set targets for 2030 to source 10% of its annual consumption from domestic extraction, 40% from processing, and 25% from recycling, while reducing reliance on any single third country to 65% or below. On July 10, 2025, the U.S. Department of Defense signed a 10-year agreement with a domestic producer, establishing a price floor of $110 per kilogram for neodymium-praseodymium oxide and purchasing $400 million in convertible preferred stock.

The export controls unfolding right now and the responses to them are an ongoing story. We have no intention of predicting who will win here. Instead, let's pinpoint four places to watch.

First, whenever you see a number, look at which axis it is measuring. Mining share and separation share are different, and the answer changes depending on whether it is oxide equivalent or metal basis, all seventeen elements or just the few used for magnets. A percentage without three vital qualifiers—the reporting institution, the point in time, and the scope of coverage—can hardly serve as a basis for sound judgment.

Second, look at who agrees to process the residues, and where. A hundred years of this story has turned on that single question. Much like Malaysia's licensing conditions, how and in which country radioactive residues are managed determines where facilities are built.

Third, watch how long price-supporting mechanisms can last. Price floors or long-term offtake agreements are tools designed to keep facilities running even after prices crash. The Congressional Research Service notes that once supply expands, prices typically revert to near the industry's marginal cost of production. In the previous cycle, there was no safety net in place. This time, one has been laid down, and seeing how long it holds is the key thing to watch.

Fourth, keep your eyes on the heavy rare earths and magnet stages. While mining shares have already come down, separation and magnet production have barely budged. A single line in a table showing that 100% of the terbium, holmium, and lutetium compounds and metals imported by the United States come from China defines where this industry stands right now.

Finally, here is the one sentence this chapter leaves behind: A map that counts where deposits lie and a map that counts the factories that separate them are two different maps. Just checking which map is shown in the news will make the exact same headline read completely differently.

If you condense it into a single sentence, it is this: This story is not about what is buried where, but about who decided to take on the work of separating those seventeen elements, one by one.