Highlights
- China currently controls 85-90% of rare earth oxide and magnet production, creating a critical global dependency.
- The United States and allies are pursuing a 'Warp Speed' strategy to develop independent rare earth magnet capabilities by 2040.
- Achieving 50% rare earth magnet independence requires:
- Significant investment ($50-75B)
- Synchronized industrial policy
- Strategic prioritization of magnet production
If China is the factory of rare earths, the West is still waiting in the loading dock. Roughly ~85–90% of refined rare-earth oxides and NdFeB magnets are born, bathed, or finished in the People’s Republic. That’s not a statistic; it’s a dependency. It means the permanent magnets that make EVs purr, wind turbines turn, missiles steer, drones hover, robots flex—and the whole clean-energy and defense economy hum—are overwhelmingly Chinese.
Washington has finally put a price on that reliance. The Pentagon is now paying a premium to accelerate U.S. NdPr (neodymium-praseodymium) capacity—well above China’s spot levels—and has committed major grant/contract funding to MP Materials. MP, the United States’ lone producing rare-earth mine, has also secured large commercial financing and a dedicated heavy-rare-earth program. Apple has committed significant multi-year collaboration and funding to U.S. magnet supply and recycling, including work with MP. None of this is charity; it’s triage.
The strategic question: How fast can the U.S. and its allies get to 50% independence—defined here as half of the world’s refining and magnet making located in the U.S., Europe, Australia (and close partners)—and how much will it cost?
What follows is the cleanest read on the runway from 2025 to 2040, separating the lights (Nd, Pr dominate magnets) from the heavies (Dy, Tb for high-temperature performance), and distinguishing oxides from metals from magnets—because confusing them is how rosy press releases become unhappy capex.
2025: The Reality Check
Supply today.
Global REO demand is on the order of ~200,000 tpa. The West—principally Lynas and MP Materials—refines only a thin slice of that, roughly single-digits to low-teens percent. Lynas is the standout: the largest ex-China separator, scaling toward ~10–12 ktpa NdPr-equivalent capacity and making early but material inroads into dysprosium/terbium separation. MP Materials is commissioning U.S. NdPr separation and a first, modest magnet line. Most other Western efforts are piloting, permitting, or pitching.
Magnets today.
China makes ~85–90% of NdFeB magnets; Japan is the clear runner-up; the U.S. and Europe together are a rounding error. America’s industrial base for metal/alloy making and sintering is almost a blank page. Recycling? Despite the headlines, on the order of ~1% of magnets are produced from recycled feed today. That Apple/MP collaboration is the beachhead, not the beach.
Heavies today.
The Achilles’ heel is unambiguous: HREEs (Dy, Tb) for high-temperature magnets remain almost entirely China-processed—with ionic-clay feed from China and Myanmar doing the heavy lifting. Early Western heavy-separation plants are coming, but not yet biting into market share.
Bottom line:
At the dawn of 2025, the West is ~90% dependent across the midstream and downstream. That’s why DoD is paying a premium and why investors are rediscovering patience as a virtue. And this is why a more comprehensive, synchronized industrial policy is needed among the West, led by America.
REEx perspective.
Among ex-China processors, Lynas ranks #1 on actual throughput and planned expansions; MP Materials ranks just behind on integrated potential but with less current oxide throughput; Iluka and Carester look like credible heavy-leaning entries moving from paper to steel over the next few years, along with a few others.
2030: The First Climb—If We Keep at It
Industry forecasts suggest 2030 will be the capacity-establishing phase. EVs and wind and a host of other verticals will have pushed magnet demand sharply higher; the world will want ~260,000+ tpa REO and ~400,000 tpa of NdFeB magnets. Two paths:
Conservative “deal-by-deal” trajectory.
If today’s announced projects land on time, Western refining could reach 50–60 ktpa—roughly 20–25% of the world's needs. The heavy lifting comes from Lynas' expansions; MP’s U.S. NdPr separation; Arafura’s Nolans project; Iluka’s Eneabba refinery (with meaningful heavy capability); REEtec (Norway), SRC (Canada), and Energy Fuels inching up from pilot to early commercial. Assets in Africa and Brazil show promise.
China still supplies ~75–80% of refined oxides in this case. On the lights vs. heavies split, Western NdPr output grows; heavy independence still lags badly—maybe <10% of global Dy/Tb separated outside China by 2030. On magnets, progress is modest: a U.S. line here, an EU pilot there, Japan steady—perhaps 12–15% of global magnets are Western/Japanese by 2030. The U.S. moves from zero to something; Europe proves it can make thousands of tonnes, not tens.
Accelerated “Operation Warp Speed” ($50B+).
With a crash program, the West can push 80–100 ktpa refining by 2030 (30–40% share). That means parallel funding of multiple Lynas/MP-scale trains and pulling U.S., EU, and Canadian projects forward with grants, loan guarantees, offtakes, and force-multiplying procurement, not to mention other policy levers such as workforce development. The money must go downstream too: build dozens of magnet lines across the U.S. and EU; lure Japanese/Korean champions to joint-venture; fund Western metal/alloy capacity in lockstep. Result: 100–120 ktpa magnets (25–30% share). China is still dominant—but parity becomes imaginable.
REEx processor table, Aug 2025 (building blocks of 2030 math).
Indicative planned capacities: Lynas ~12,000 tpa LREE and ~3,000 tpa HREE; MP Materials ~6,000 tpa LREE separation; Iluka ~5,500 tpa LREE + ~725 tpa HREE; REEtec ~1,000 tpa; Ucore ~5,000 tpa planned.
2035: The Inflection—or the Stall
By the mid-2030s, the story turns on execution, not intent.
Conservative case (30% refining; ~15–20% magnets).
The West could be producing ~100 ktpa REO—about 30% of demand—as Arafura fully scales out, Lynas debottlenecks again, MP adds another train, and a handful of North American/African projects finally ship oxides. Europe’s CRMA ambitions materialize in _pockets_—Solvay, Neo, and one or two new plants. Respectable, not decisive. Even with Iluka online and a couple of ionic-clay projects outside China, the West might cover a third or less of Dy/Tb by 2035. Recycling begins to nibble at the problem, contributing ~5–10% of magnet feed at best. On magnets, a functional Western base delivers ~100 ktpa (15–20% share). It’s progress, but the awkwardness remains: oxides are Western; too many magnets are still made offshore.
Warp-Speed case (~50% refining; ~40–50% magnets).
With sustained money and political will, Western/allied REO capacity reaches ~180–200 ktpa—half the market. That implies a dozen-plus new mines/refineries beyond today’s pipeline and serious HREE separation standing up in France, Australia, and North America. Could players like Phoenix Tailings disrupt the model, with separation and refining breakthroughs based on recycling?
Non-ChinaDy/Tb supply could cover 30–40% of global needs by 2035 via Iluka, Carester, a ramped-up MP (with feedstock secured), selective ionic-clay developments, and heavy-doped scrap recycling. The decisive leap is ~250 ktpa magnets (40–50% share): 10–15 large factories across the U.S., EU, and partner nations, synchronized with metal/alloy expansion. Government procurement (defense, fleets, grid) and auto/wind offtakes keep utilization high and financing cheap.
HREE wild cards. Among ex-China processors, Iluka and Carester could emerge as the heavy levers—small in tonnage, giant in strategic leverage. If they slip, the whole heavy-magnet equation slips with them.
2040: Choose Your Adventure
By 2040, global magnet demand could top ~800,000 tpa, with REO needs ~400–450 ktpa. Two outcomes:
If we drift: a larger, still-dependent West.
Refining inches to 150 ktpa REO (33%). Better than today, not enough to claim “independence.” Magnets reach 200 ktpa (25%)—China still carries ~75%. Heavies become a meaningful minority sourced outside China, but Dy/Tb are still largely price-set and volume-set in the Chinese sphere.
If we surge: fifty percent—for real.
Western/allied refining hits ≥250 ktpa (≥50%). The West operates large LREE refineries and multiple HREE hubs processing clay-derived and recycled streams. Magnets reach 400 ktpa (50%), spread across the U.S., EU, Japan, Korea, Australia, India—fed by an integrated allied metal/alloy base. Recycling delivers 10–15% of magnet feed—especially valuable for heavies. The magnet-grade REO market alone could be $40–50B/year by 2040; finished magnets likely >$100B. Keeping half inside allied economies permanently changes the calculus on supply security and industrial policy.
The Friction You Can’t Ignore
| Factor | Implication | Mitigation |
|---|---|---|
| China can (and will) play price hardball | Underpricing (“dumping”) can bankrupt fragile rivals | Without long-term offtakes, price floors, stockpiles, or tariff backstops, expect déjà vu. |
| Permitting is policy | In the U.S. and EU, a rare-earth project can spend a decade moving paper. Thorium tails spook communities; solvent-extraction plants draw scrutiny. | “Warp Speed” isn’t just money; it’s time. |
| Heavies are a different sport | China spent decades mastering Dy/Tb cascades. Western alternatives (RapidSX, ion exchange, chromatography) are promising but unproven at thousands of tonnes. | Carester and Iluka are vital; without them (and successors), you have a magnet body without a high-temp soul |
| Don’t overbuild oxides and underbuild magnets | It’s easy to crow about “capacity” at the oxide stage and discover your metal/alloy and magnet shops are still drawings | The order of operations must be synchronized: mine → separate → metal/alloy → magnet, all backed by binding demand. |
| Geopolitics won’t sit still | Export controls, new alliances, and tech pivots (e.g., Dy-lean designs or alternate motors) can scramble demand and pricing. | Diversity—of ores, processes, geographies, and customers—is insurance. |
The Math of Independence
- Target (2040): ≥50% of REO refining and NdFeB magnet production done in the U.S./EU/Australia/allies.
- Volumes: ~200–225 ktpa REO and ~400 ktpa magnets on Western soil.
- Capital: The conservative path inches there after 2040—if at all. A “Warp Speed” push ($50–$75B blended across mines, midstream, metals, magnets, recycling over ~15 years) can hit it by 2040.
- Operating reality: Unit costs will be higher at first. That’s why premium NdPr pricing and grants are a feature, not a bug—they buy time to scale and learn. The payoff is resilience and bargaining power when the next export control, price crash, or geopolitical storm hits.
The Playbook (No Mystery, Just Discipline)
Stack the deck for magnets.
Prioritize metallization and magnet lines equal to or ahead of oxide capacity. Offer 10-year offtakes (defense, fleets, grid), bonus points for Dy-lean designs and recycling integration.
De-risk heavies.
Back Iluka- and Carester-class plants; fund at least two more heavy hubs by 2030 perhaps in Brazil, Australia? Launch a Dy/Tb stockpile and prize grants for proven heavy-recovery from scrap.
Shorten the permit.
Create “National Security Critical Materials Corridors” with single-window approvals and known timelines; tie funding to community-benefit agreements and transparent waste plans.
Make prices boring.
Use price floors, counter-dumping tariffs, and a strategic magnet reserve to stabilize new entrants through cycles. Remember when real markets emerge contacts become more standardized.
Ally-source and ally-sell.
Australian/Canadian ore → U.S./EU separation → U.S./EU/Japan/Korea metals and magnets. Contractually fence it to avoid leakage back to the Chinese midstream.
Measure what matters.
Publish quarterly oxide → metal → magnet conversion ratios, not just “nameplate capacity.” Glamour lives at the magnet, not the mine.
The Verdict
Can the West break the 90% habit? Yes. On the current, careful path, we reach ~30% refining and ~25% magnets by 2040—safer than today, still vulnerable. On the Warp-Speed path, with real money and more integrated industrial policy, including real permitting reform, we can hit ≥50% in both refining and magnets by 2040, with NdPr essentially solved by the mid-2030s and Dy/Tb no longer a single-point failure.
The choice isn’t markets versus industrial policy; it’s dependency versus pricing power. The former is cheap until it’s not. The latter costs more now and saves you later, and now has a security premium impact—on balance sheets, on carbon, and when the next shock comes. If you want a different ending than the Molycorp saga, write a different second act.
And remember the quiet tell in every supply chain: the value—and the leverage—live at the magnet. Build there first, and the rest of the chain will stop feeling like a bet and start behaving like a business.
Source note: Company disclosures and REEx’s August 2025 Rare Earth Processor Rankings underpin the ex-China capabilities cited (e.g., Lynas, MP Materials, Iluka, REEtec, Ucore; current vs. planned LREE/HREE capacity). The rankings synthesize stage, throughput, technology, feedstock flexibility, offtakes, and government support—useful for gauging who can realistically move tonnage outside China.
0 Comments