Top Rare Earth Stocks 2026: Securing the West's Supply

Rare earths technically aren't rare.

Cerium is more common in the Earth's crust than copper. What is rare is the ability to dig the stuff up, crack it, separate 17 nearly identical elements from each other, turn four of them into metal, and press that metal into a magnet the size of your thumbnail that can spin a Tesla motor at 20,000 rpm.

China figured that out 30 years ago. Everyone else is figuring it out right now, at gunpoint.

The four elements that matter are neodymium, praseodymium, dysprosium, and terbium.^[1] Put them in a magnet and you get the strongest permanent magnet humans know how to make. Every EV, every wind turbine, every F-35, every humanoid robot, every hard drive in an AI data center runs on them. Pull the magnets and the machine is a paperweight.

In April 2025, China put export licenses on seven rare earths. In October it went further, restricting anything with even trace Chinese content. Then in November it "suspended" the October rules for a year after a Trump-Xi handshake. But in 2026 it kept shipping exactly zero dysprosium, terbium, and yttrium to Japan for months on end. Terbium outside China hit $4,500/kg. Inside China it was under $1,000. Yttrium went to $1,100+ outside China versus about $10 domestic.

Put simply, rare earth elements are now geopolitical leverage.

Washington's answer was blunt. The Department of War bought 15% of MP Materials, guaranteed it $110/kg for NdPr for ten years, and promised to buy every magnet out of a factory that doesn't exist yet. Then it did versions of the same deal with Lynas, Serra Verde, Vulcan Elements, ReElement, and Sunrise. Fifty-four countries showed up to a Critical Minerals Ministerial in February. The US launched a strategic minerals reserve modeled on the oil one.

There's no longer one rare earth price. There's the Chinese price, and there's the price the West will pay to never depend on it again.

The ex-China premium is structural, not a trade-war blip. Governments are writing $110/kg floors into ten-year contracts. Contracts don't get un-signed when a customs office reopens. But most of the companies in this sector are still building, not producing, and building rare earth plants is where money goes to disappear.

We're slicing this theme by where a company sits on the road from rock to magnet:

• Producing Now – already selling separated oxide, with government floors under them
• Mine-to-Magnet Builders – assembling the full chain, some of it working today
• The Next Wave – funded developers with a construction start date you can put on a calendar
• Picks-and-Shovels – technology that processes other people's ore
• Private Bellwethers – the magnet makers Washington is bankrolling that you can't buy yet

These three sell separated rare earth oxide today. Each has a government contract that guarantees a minimum price, so if Beijing floods the market next spring, they keep getting paid.

MP Materials NYSE: MP

If you only know one name in rare earths, it's this one. MP Materials owns Mountain Pass in California, the only commercial rare earth mine in the United States, good for roughly 13% of global mine output. It stopped shipping concentrate to China last year. Now it separates its own NdPr oxide and sells to Japan and Korea, with 1,006 tonnes sold in Q2 (up 127% year on year) and $126M of revenue.

The July 2025 Pentagon deal is the whole story. $400M in preferred stock. A $110/kg NdPr floor for ten years, with the government taking 30% of any upside above it. A promise to buy 100% of the magnets from a 7,000 tonne "10X" plant in Northlake, Texas, targeting 2028. Apple added $500M for recycled magnets from 2027. The heavy rare earth circuit at Mountain Pass finished mechanical completion in May and should produce first US dysprosium and terbium later this year. At a ~$10B market cap it's priced like the winner. It probably is. The question is whether you want to pay for 2028 today.

Lynas Rare Earths ASX: LYC

Lynas is the largest rare earth producer outside China, and the one with the longest track record of actually doing this. Mt Weld in Western Australia feeds a separation plant in Malaysia that's been running since 2012. In May and June 2025 it became the first company outside China to commercially separate dysprosium and terbium. Samarium followed in March 2026.

The contracts are the moat. Japan's JOGMEC and Sojitz extended their offtake to 2038 at a $110/kg NdPr floor, taking 5,000 tonnes a year plus half of Lynas's heavy rare earth output. The US Department of War signed its own ~$96M letter of intent at the same floor. In July, Lynas put A$50M into Korea's JS Link to build a 3,000 tpa magnet factory next door in Malaysia. Two warnings: the Texas plant is dead after permitting fights, and the new heavy rare earth facility just got re-costed 63% higher. Lynas executes, but not cheaply.

Energy Fuels NYSE American: UUUU

Energy Fuels runs the only conventional uranium mill in America, White Mesa in Utah, and quietly turned it into the only US facility licensed to make high-purity rare earth oxides from monazite at commercial scale. Phase 1 does up to 1,000 tonnes of NdPr a year. It made 99.9% dysprosium oxide in August 2025 and terbium in March 2026, the first primary US heavy rare earth production in decades.

Then in June it agreed to buy VAC, the German magnet maker with plants in Hanau and Sumter, South Carolina, for $1.9B. Close it in early 2027 and Energy Fuels becomes mine-to-magnet overnight. Feedstock comes from Chemours mineral sands in Florida and Georgia, the Donald project in Australia, and its own deposits in Brazil and Madagascar. The stock is down 30% over six months, mostly because the $700M convertible plus a $1.9B acquisition made shareholders nervous about dilution. Two commodities, one balance sheet, a lot of moving parts.

These two are assembling the full chain, and parts of it already work. They're further along than the developers, but they don't have MP's scale or Lynas's twelve-year operating history.

USA Rare Earth NASDAQ: USAR

A year ago USA Rare Earth was a Texas deposit with a PowerPoint. Now it's a $4.4B company with a running magnet line in Stillwater, Oklahoma (commissioned March, targeting 600 tpa by Q4), a metal-and-alloy business in the UK via the Less Common Metals acquisition, a second magnet campus breaking ground in South Carolina with up to $1.6B in Commerce Department support, and a Colorado demo plant that produced commercial-grade dysprosium and NdPr oxide from recycled magnet swarf in July.

The big swing is Serra Verde. USAR agreed in April to buy the Pela Ema mine in Brazil for about $2.8B in cash and stock. Pela Ema is the only mine outside Asia producing all four magnet rare earths at commercial scale, with over $1B already sunk into it. The Department of War put $750M into a special purpose vehicle that buys 100% of Phase 1 output, plus a $300M forward purchase and a $500M bank revolver. Shareholders voted yesterday, August 28, and the company expects to close promptly after. Stock is up ~60% this year. Q2 revenue was $5.8M. You're paying for the story, not the income statement.

Neo Performance Materials TSX: NEO

Neo is boring, and boring is the point. Neo is the only Western-listed company running both rare earth separation and sintered magnet manufacturing at commercial scale outside China, both in Estonia. It's also one of the only rare earth companies on this list that makes money: 2025 adjusted EBITDA guidance was $67M to $71M.

The Narva magnet plant opened in September 2025 at 2,000 tpa, scaling to 5,000, with Schaeffler and Bosch already qualifying product. It's ramping toward mass production this half. At $1.1B market cap, Neo trades at a fraction of MP or USAR for a business that already does what those two are promising to do. The catch: it's Canadian-listed, Europe-focused, and hasn't gotten the Pentagon love that lifted the US names.

Developers with real money in the bank and construction dates you can check. None of these produce yet. All of them could double or halve on a permit, a cost re-estimate, or a financing close.

Arafura Rare Earths ASX: ARU

Arafura took final investment decision on the Nolans project in May. Construction starts in September. That sentence alone puts it ahead of 90% of rare earth developers. Nolans in the Northern Territory would produce 4,440 tonnes of NdPr a year, about 4% of world supply, for 38 years.

The money is in the bank: $775M in export credit debt, A$200M in government convertible notes, A$475M raised last October, another A$350M in July, and A$723M cash at June 30. Hancock Prospecting (Gina Rinehart) owns about 17.5%. Five binding offtakes including Hyundai/Kia and Siemens Gamesa, with the two newest priced off a seaborne index instead of China's. Total funding need is about $1.9B. The stock has been diluted hard to get here. Now it has to build the thing.

Iluka Resources ASX: ILU

Iluka is Australia's biggest mineral sands miner, and it's spending A$1.7B to A$1.8B on Eneabba, which it calls Australia's first fully integrated rare earth refinery and the only one designed from day one to separate both light and heavy rare earths. The Australian government lent it A$1.65B, the largest critical minerals loan in the country's history. Commissioning is set for 2027.

Iluka already has the feedstock. Decades of monazite sat in a stockpile as a byproduct of its main business. So you get a producing mineral sands company at a $2.2B market cap with a rare earth refinery bolted on. The risk is that novel refineries rarely commission on schedule or budget, and the stockpile runs out eventually.

Aclara Resources TSX: ARA

If you believe the heavy rare earth squeeze (dysprosium, terbium) is the real trade, Aclara is the purest bet on it. Its Carina project in Brazil is ionic clay, the same deposit style China's own heavy rare earth supply comes from, and in 2025 became the first ionic clay project anywhere to declare compliant reserves. The March feasibility study targets 156 tonnes of dysprosium and 27 tonnes of terbium a year. Its water-based extraction avoids acid leaching and radioactive waste, which matters for permits.

Downstream it's planning a $277M separation plant in Louisiana with groundbreaking in Q4 and production around 2028, sized to supply most US dysprosium and terbium demand for EVs. A $50M placement in May brought in New Hartsdale Capital as the largest shareholder. Early works start mid-2026; full construction is 2027. At $698M market cap it's the smallest name here and the most levered to heavy rare earth prices going either way.

One name. The separation step is where China's chokehold is tightest, and Ucore sells a machine that does it.

Ucore Rare Metals TSXV: UCU

Ucore is a separation technology company, not a miner. Its RapidSX platform is a continuous-flow solvent extraction system it says is smaller, faster, and cheaper to build than the conventional mixer-settler trains China uses. The business model is to process other people's feedstock for a fee.

The Department of War has put $22.4M into Ucore's Strategic Metals Complex in Alexandria, Louisiana, and gave it a DPAS priority rating, which forces suppliers to bump Ucore to the front of the line. First commercial machine, about 600 tpa, is targeted for H1 2027 inside a facility designed for 9,600 tpa across three lines. It has a feedstock deal for Australian monazite with Wyloo and Hastings, and Canadian government approval for up to C$36M. At $225M market cap this is a technology bet with a government tailwind. If RapidSX works at scale, every developer on this list is a customer.

The magnet end of the chain is where Washington is writing the biggest checks, and the recipients are mostly private. You can't buy them. You should still watch them, because they set the prices the public names get.

Vulcan Elements (Private)

Vulcan is a North Carolina startup that makes sintered NdFeB magnets for defense and commercial customers, and in November 2025 it landed a $1.4B package: a $620M Office of Strategic Capital loan, $50M of equity from the Commerce Department, and $550M from private investors, to build a 10,000 tpa US magnet plant. The Pentagon holds warrants. Once that plant runs, it's the largest magnet factory outside Asia and the benchmark USAR and MP's 10X plant get measured against. No IPO talk yet, but a government-backed unicorn doesn't stay private forever.

Noveon Magnetics (Private)

Noveon is the one that's already shipping. Its San Marcos, Texas plant makes sintered NdFeB magnets from recycled end-of-life magnets, and in January it raised a $215M Series C led by OneIM, the fund of ex-SoftBank Vision Fund chief Rajeev Misra (also an early Vulcan backer). Congress banned Chinese magnets in Department of Defense systems starting 2027. Noveon is the company most ready to fill that order today.

ReElement Technologies Private, subsidiary of NASDAQ: AREC

ReElement is the refiner in Vulcan's chain. It uses Purdue-licensed chromatography to purify rare earth oxides from recycled magnets, e-waste, and mined concentrate at a campus in Marion, Indiana, and got $80M of the Vulcan package. In May, POSCO International signed a JV to build US separation capacity with it. ReElement is owned by American Resources Corp, a small NASDAQ coal-turned-critical-minerals company. There's talk of a ReElement IPO in late 2026. If that happens, AREC shareholders find out what the subsidiary was worth. If it doesn't, AREC keeps burning cash.

Failure mode one: China opens the tap. Every developer's feasibility study assumes NdPr somewhere north of $110/kg. China's domestic NdPr benchmark already slipped below that floor in August as production quotas ran ahead of demand. If Beijing decides it would rather bankrupt Western projects than restrict exports (it did exactly this in 2011-2015 and again in 2023-2024), spot could fall to $60 and stay there. The producers with floors survive. Everyone without a contract doesn't. Watch two things: China's monthly customs data on dysprosium, terbium, and yttrium to Japan (still zero as of June), and November 10, 2026, when the one-year suspension of the October 2025 controls expires. If controls come back, the premium widens. If they quietly lapse and volumes normalize, the whole "ex-China premium" thesis gets marked down.

Failure mode two: the concrete doesn't get poured. Rare earth processing is chemistry, and every deposit has different chemistry. Flowsheets don't transfer. Lynas re-costed a plant by 63%. Its Texas project died on a wastewater permit. Iluka's refinery has drifted up in cost. Evolution Metals (NASDAQ: EMAT), which we deliberately left off this list, listed in January at $2B, has going-concern warnings in two straight filings, and is down 88%. The tell is always the same: capex re-estimates, FID dates that slide a quarter and then another, and equity raises at lower and lower prices. Arafura, Viridis, and Aclara all have final investment decisions or construction starts in the next four months. Those are pass/fail exams.

The next 12-18 months are when announcements have to turn into tonnes.

Fall 2026. USA Rare Earth's Serra Verde deal closes and Pela Ema ramps toward 4,000 tpa TREO by year-end, with first deliveries to the government SPV early in Q4. Arafura breaks ground at Nolans in September. Viridis (ASX: VMM) targets FID in Q4. Aclara breaks ground in Louisiana. MP Materials should announce first dysprosium and terbium out of Mountain Pass. USAR's Stillwater line hits 600 tpa.

November 10, 2026. China's export control suspension expires. This is the single biggest binary in the sector. Either Beijing re-imposes the trace-content rules and every ex-China name gaps up, or it extends the truce and the heavy rare earth premium starts to compress.

Early 2027. Energy Fuels closes VAC and becomes mine-to-magnet. The Pentagon's ban on Chinese magnets in defense systems takes effect, making Noveon, Vulcan, VAC's Sumter plant, and MP's Independence facility the only qualified suppliers. Ucore's first RapidSX machine comes online in Louisiana. Iluka begins commissioning Eneabba.

Wildcards. A ReElement IPO. A Vulcan IPO. Another sovereign equity stake, since Washington now owns pieces of MP, Vulcan, ReElement, and the Serra Verde SPV. And whether the NdPr price the West pays finally decouples from the price China reports.

The producers are expensive because they're the ones with the contracts, and the contracts are the asset. The developers are cheap because most of them won't make it. If you're buying the theme, own a producer for the floor and one developer for the torque, and read the construction updates.

Not investment advice. Rare earth stocks can lose half their value on a customs bulletin. Size accordingly.