Rare Earths in 2026: China's dominance tightens as the West scrambles for alternatives

The rare earths market in mid-2026 captures a fundamental tension: China's grip on the supply chain remains as strong as ever — roughly 60% of global mining and an overwhelming ~90% of separation and refining capacity — while Western governments pour unprecedented capital into building alternatives from scratch. The result is a market that is structurally tight, geopolitically charged, and increasingly bifurcated between a Chinese-dominated commodity flow and a Western subsidized parallel chain.

Prices correct from Q1 highs, but structural deficit intact. NdPr oxide hit $99.61/kg on the Shanghai Metals Market benchmark on May 1, down 21% from its March peak of ~$126/kg, pulling back from a staggering 105% rally in the first quarter of 2026. The correction reflects profit-taking and inventory adjustment rather than any improvement in underlying supply-demand balances. The structural supply deficit in NdPr — the critical blended feedstock for NdFeB permanent magnets — remains intact, with global mine supply still lagging projected demand growth of 7–8% per year for magnet-grade rare earths.

Dysprosium oxide has proven more resilient, holding at $177/kg domestically (China) and averaging $292/kg FOB, supported by structurally tied demand from high-temperature magnet grades in EV traction motors and wind turbine generators. Terbium oxide corrected 18.6% to ~$886–895/kg, but heavy rare earth supply is even more constrained than light, with Lynas only now achieving the first commercial separated dysprosium outside China in May 2025, and terbium following shortly after.

China's control strategy evolves beyond quotas. China no longer relies on WTO-disputed export quotas. Instead, Beijing deploys a sophisticated suite of tools: an outright ban on rare earth extraction, separation, and smelting technology exports (December 2023); targeted export controls on medium and heavy rare earths (April 2025); and a sweeping October 2025 package covering materials, processing equipment, and related technologies — including extraterritorial provisions that assert control over foreign-made products using Chinese technology.

However, on November 7, 2025, China suspended the October 2025 controls for one year through November 10, 2026, following the Trump-Xi meeting in South Korea. This suspension is temporary and reversible — the 2023 technology export ban remains, as does the blanket prohibition on dual-use exports to US military end-users. The message is calibrated: China can tighten or loosen supply at will, and the sword of Damocles hangs over every Western supply chain initiative.

MP Materials achieves the first US mine-to-magnet chain. The most significant Western milestone came in December 2025, when MP Materials shipped its first commercial NdFeB permanent magnets from its Independence Facility in Fort Worth, Texas — the first US-made rare earth magnets in over two decades. MP produced 2,599 metric tons of NdPr oxide in FY2025, up 101% year-on-year, with a run-rate target of 6,000 t/year by end of 2026. The company also redirected all Mountain Pass output away from China to US, Japanese, and South Korean supply chains.

A heavy rare earth separation plant at Mountain Pass is targeting mid-2026 commissioning with ~200 t/year of dysprosium and terbium capacity. The longer-term vision is the 10X magnet campus in Northlake, Texas — a $1.3 billion facility adding ~10,000 t/year of NdFeB magnet capacity, backed by a 10-year Department of Defense offtake agreement and a $110/kg NdPr oxide price floor.

Lynas scales as the only non-Chinese heavy REE processor. Lynas Rare Earths produced 6,558 t of NdPr in FY2025, up 16% YoY, targeting ~10,500–12,000 t/year at full ramp from its Mt Weld mine and Kalgoorlie processing facility. The Kalgoorlie plant's ramp has been hampered by power disruptions, but the strategic prize is Lynas's position as the first commercial separator of dysprosium and terbium outside China — a capability that US DoD has underwritten with ~$258 million in funding for a Texas heavy REE processing facility.

Magnet demand accelerates on EV and wind deployment. Global NdFeB permanent magnet demand is projected at $28.34 billion in 2026, up 7.1% from 2025, reaching $37.67 billion by 2030. Permanent magnet applications — which account for 31.2% of total rare earth demand — are growing at a 7.4% CAGR through 2031, driven primarily by EV traction motors and offshore wind turbine generators. Each EV requires 1–2 kg of NdFeB magnets, while direct-drive offshore wind turbines consume roughly 600 kg of NdPr oxide per MW.

The IEA projects demand for magnet rare earths will increase 45% between 2024 and 2030, and Adamas Intelligence forecasts a cumulative NdFeB undersupply of 246,000 tonnes by 2040. Even with thrifting technologies and recycling (projected to supply up to 10% of demand by 2030), the magnet supply gap will widen before Western processing capacity catches up.

US-EU Critical Minerals Partnership adds policy backbone. In 2026, the US and EU signed a formal critical minerals partnership under a new MoU, explicitly aimed at loosening China's grip on rare earth supply chains. The agreement covers coordinated trade policies, potential border-adjusted price floors, offtake agreements, investment screening coordination, and even stockpiling cooperation. Combined with the Defense Production Act's deployment of over $150 million into rare earth processing, $258 million into Lynas's Texas facility, $400 million in equity and $150 million in loans into MP Materials, and $94 million into E-VAC Magnetics for domestic NdFeB magnet manufacturing, the Western policy apparatus is now backing rare earth supply chain independence with scale.

Yet the gap remains daunting. China still produces roughly 138,000 tonnes of NdFeB magnets annually versus MP's initial ~1,000 t/year capacity. The 10X campus, when commissioned in 2028, will add 10,000 t/year — still less than 8% of China's current capacity. The West has the capital and the political will, but building a parallel rare earth supply chain at scale is measured in years, not quarters.