The Lyceum: Critical Minerals Weekly — Mar 25, 2026

This is the week the calendar started biting. China's March 25 Beijing briefing on rare earth licensing lands today, the Pentagon's bid window for critical mineral processors closed March 20, DOE letters of intent for $500 million in battery processing money are due Thursday, March 27, and the January 1, 2027 defense magnet policy is now nine months out with almost no domestic oxide-to-metal capacity online. Meanwhile, Beijing isn't just restricting exports — it's choosing who gets them, routing gallium to Germany while cutting Japan off entirely. The money is finally flowing, the deadlines are real, and the question is no longer whether the West will build alternative supply chains but whether it can do so before the clocks run out.

• Lynas–JARE NdPr Price Floor Agreement (Lynas / Japan Australia Rare Earths): Fixed $110/kg NdPr floor with upside-sharing above $150/kg, effective through revised long-term supply arrangement.
• FORGE Critical Minerals Initiative (U.S. State Department): Replaces the Minerals Security Partnership with a project-finance-steering forum designed to channel allied capital to specific midstream builds.
• DOE $500M Battery Processing NOFO (Department of Energy): Funding opportunity open for processing, cathode/anode manufacturing, and recycling — LOIs due March 27, full applications April 24.
• NSF Tech Metal Transformation Challenge Winners (National Science Foundation): Eight teams selected from 130 to prototype e-waste-to-critical-metal recovery systems, with Naval Research Laboratory collaboration.
• Canada Critical Minerals Production Alliance — Round 2 (Natural Resources Canada): $12.1 billion in new deals across 30 partnerships spanning mining through refining.
• REalloys Saskatchewan Heavy Rare Earth Metallization Plant (REalloys Inc.): Fully financed buildout for oxide-to-metal conversion of dysprosium and terbium — the first such facility outside China.
• DIB Consortium Critical Minerals RPP Closed (Defense Industrial Base Consortium): Title III Request for Project Proposals for domestic processing closed March 20 after a three-week window.

Forget the binary question of whether China's export tap is open or closed. The new question is who Beijing pours for.

Trade-flow data from Tradium GmbH shows China has effectively zeroed out gallium and germanium shipments to Japan in early 2026, while routing the bulk of gallium exports to Germany. This is precision allocation, not blanket restriction — and it reframes supply risk from a commodity problem into a political-access problem. Gallium goes into compound semiconductors, 5G infrastructure, and next-generation power electronics. Germanium is critical for fiber optics and infrared military systems. Japan, one of the world's largest consumers of both, is now functionally cut off from its primary supplier.

If this pattern holds, Japan may accelerate domestic gallium recovery (from zinc refining byproducts) or lock in allied sourcing deals at premium prices. The observable signal: watch whether JOGMEC announces emergency gallium procurement contracts or whether Japanese semiconductor fabs begin qualifying non-Chinese gallium sources in Q2. If neither happens by summer, Japanese chipmakers face real input constraints by year-end.

Today's CCCMC industry briefing in Beijing — scheduled for March 25 — is the channel where licensing or enforcement changes get telegraphed. If new HS codes or tighter end-use requirements emerge, traders will pre-emptively reroute cargoes within days.

If you've been tracking the defense magnet supply chain, you know the bottleneck isn't mining or even separation — it's the step where separated rare earth oxides become actual metals and alloys that magnet manufacturers can use. China controls that step almost entirely.

REalloys Inc. announced a fully financed buildout of a heavy rare earth metallization facility in Saskatchewan. The plant will convert separated oxides of dysprosium and terbium — the elements that let permanent magnets work at the extreme temperatures inside jet engines and missile guidance systems — into finished metals and alloys. This directly targets the federal policy taking effect on January 1, 2027, restricting Chinese-sourced rare earth magnets in U.S. military platforms.

If REalloys hits its ramp schedule, it becomes the single most important de-risking event for the Western magnet supply chain this year. If construction slips or yields disappoint, defense primes will face the 2027 restriction with no compliant source of heavy rare earth metals at commercial scale. The signal to watch: whether Lockheed Martin, RTX, or Northrop Grumman sign offtake agreements with REalloys by Q3 — that would confirm the plant's output meets defense-grade specifications.

On March 20, the Defense Industrial Base Consortium closed a three-week Request for Project Proposals under Defense Production Act Title III, seeking partners to build or expand U.S. processing capacity for a named list of strategic materials. The call targeted chemical conversion, refining, and metal-making — midstream, not mines.

That short fuse is a tell. The Department of Defense already has a universe of candidates and is running a selection race, not a discovery process. Budget justification documents list specific materials: arsenic, bismuth, gallium, graphite, hafnium, niobium, titanium, and several rare earths. Multi-metal processing platforms — think niobium-scandium-titanium-REE projects — will likely screen better than single-commodity bets.

If awards land by late Q2 with Title III funding attached, expect the first government-backed midstream plants to break ground before year-end. If the process stalls in interagency review, the 2027 magnet restriction becomes a policy without a supply chain — and defense primes start lobbying hard for waivers.

This is getting almost no attention outside compliance circles. China's updated 2026 Import-Export Licensing Catalogue, administered by MOFCOM and MIIT, now requires export licenses for samarium, gadolinium, and lutetium compounds. Samarium goes into specialty permanent magnets. Gadolinium is in MRI contrast agents and nuclear reactor control rods. Lutetium is in PET scanners and emerging cancer therapies. These are not fringe materials.

The sleeper provision: extraterritorial enforcement, with a November 2026 operative date. A factory in Germany or South Korea using Chinese-origin gadolinium could, in theory, need Beijing's permission to export its finished product. That's the logic of China's existing dual-use framework extended to rare earths — and it means 12 of 17 rare earth elements now require Chinese export licenses.

If procurement teams don't map their element-by-element exposure to the actual MOFCOM license list before November, they risk discovering compliance gaps after enforcement begins. The observable signal: watch whether European or Korean manufacturers begin qualifying non-Chinese gadolinium and samarium sources in Q2 — if they don't, they're betting that enforcement will be lax. That's a bet, not a strategy.

The Department of Energy's new funding opportunity targets the exact gap everyone keeps identifying and nobody has closed: midstream processing. The $500 million NOFO prioritizes projects that convert raw feedstocks into battery-grade lithium, nickel, and cobalt; manufacture anodes and cathodes; and scale recycling facilities. Non-binding letters of intent are due March 27 — two days from now — with full applications due April 24.

The short timeline is designed to attract shovel-ready projects, not concept papers. This is capital for commercial-scale factories producing synthetic graphite, cathode materials, and specialty metals — an explicit admission that America's supply chain problem isn't finding rock but building the factories that turn rock into battery-grade material.

If the DOE funds 8–12 projects at $40–60 million each and construction starts by early 2027, the U.S. midstream map changes materially within three years. If the money gets stuck in environmental review or spreads too thin across too many early-stage applicants, it becomes another round of announcements without tonnage. Watch which companies publicly signal LOI submissions this week — that's your shortlist of who's actually ready to build.

The most important number in rare earth finance this month isn't a spot price — it's $110 per kilogram. That's the floor price Lynas Rare Earths locked in for neodymium-praseodymium (NdPr) — the magnet material in EV motors and F-35 jets — through its revised supply agreement with Japan Australia Rare Earths. Above $150/kg, upside is shared with a cap. Below $110, Japan pays the floor anyway.

This is market design, not just a contract tweak. It makes non-Chinese rare earth refining bankable by guaranteeing returns even if Beijing floods the market. JOGMEC has reportedly extended the arrangement through 2038, effectively underwriting Lynas's project economics for over a decade.

The USTR's plurilateral price-floor consultation closed March 19, and the Lynas deal is the private-sector proof that the concept works. If the EU and Japan join formal negotiations in April — as market observers expect — allied price stability architecture moves from theory to treaty. If the consultation produces no follow-up by May, the Lynas deal stays a bilateral curiosity rather than a template. Watch for a USTR "intent to negotiate" announcement.

Rio Tinto will slow construction of the Nemaska lithium processing plant in Québec, cutting its contractual workforce roughly in half while insisting the overall production timeline is "largely unchanged." The plant is designed to convert spodumene from Whabouchi into battery-grade lithium hydroxide for North American cathode plants.

"Largely unchanged" is doing heavy lifting. Slowing construction stretches execution risk, pushes milestones into an uncertain pricing environment, and sends a signal: if a major with Rio Tinto's balance sheet is tapping the brakes on capex, juniors with similar flowsheets are on much shakier ground. Lithium carbonate has recovered to roughly $21,500/t as of March 2026 — enough to pull some brownfield expansions back into play — but not enough to make every greenfield project comfortable.

If Rio quietly revises its capex guidance or startup window later this year, expect a cascade of schedule changes across the North American lithium pipeline. If lithium stabilizes above $20,000/t through Q2 and Rio holds its timeline, the pause was a cost-management exercise, not a strategic retreat. The distinction matters for every OEM counting on North American lithium hydroxide by 2028.

The province of San Juan approved the updated environmental impact study for Josemaría, a massive copper-gold project jointly held by BHP and Lundin Mining, in under six months — remarkably fast for Latin American mining permitting. The project is expected to produce 200,000 tonnes of copper annually in its first five years.

This matters because copper has historically been the "easy" critical mineral — geographically diversified, well-understood metallurgy. It's not easy anymore. J.P. Morgan projects a 330,000-metric-ton refined copper shortfall in the U.S. this year, amplified by tariff-driven preemptive buying. Every new permitted project changes the medium-term supply calculus.

If Josemaría moves to final investment decision and construction begins on schedule, it becomes one of the most significant new copper sources for the late 2020s. If permitting stalls at the federal level or BHP and Lundin defer FID on cost grounds — the same capex pressure hitting Rio Tinto's lithium plant — the copper deficit gets worse. Watch for an FID announcement and whether Argentina's national government matches San Juan's permitting speed.

CATL and BYD are now mass-producing sodium-ion cells, and new prototypes from BAIC Group and associated Chinese developers are hitting 170 Wh/kg energy density with ~92% capacity retention in lab tests at -20°C and 4C fast charging. Sodium is roughly 450 times more abundant than lithium. The chemistry uses no lithium, no cobalt, and no nickel.

This isn't a lithium killer — energy density still trails LFP for vehicle range — but it creates a hard ceiling on lithium prices in stationary storage and entry-level EVs, the two fastest-growing battery segments. For procurement teams, it means modeling substitution risk alongside supply risk: every sodium-ion cell deployed in a grid battery is a cell that doesn't need lithium carbonate.

If sodium-ion captures even 15% of the stationary storage market by 2028, lithium demand projections built on 100% lithium-ion assumptions are materially overstated. The signal: watch whether non-Chinese battery manufacturers (Samsung SDI, LG) announce sodium-ion production lines — that's when substitution moves from Chinese advantage to global structural shift.

A Saskatchewan factory learning to turn rocks into jet-fighter magnets, a German chipmaker discovering its gallium shipments now require Beijing's blessing, and eight NSF-funded teams trying to extract the periodic table from your old laptop — the supply chain is being rebuilt one absurd bottleneck at a time. Somewhere in a MOFCOM office, a bureaucrat is deciding which countries deserve gallium this quarter, and the answer is apparently "not Japan" — which is how a Tennessee smelter gets built on the premise that server farms are the new mines. Stay paranoid.

If someone you know sources anything that plugs in, flies, or explodes, forward this to them — they need it more than they think.

The White House handed congressional leaders a six-principle AI legislative blueprint on March 20, urging states to stand down from their own rules. Read → The White House Hands Congress an AI Rulebook