A Pentagon rule taking effect January 1, 2027 bars defense contractors from delivering NdFeB or samarium-cobalt magnets, or end items containing them, whose materials were mined, refined, separated, melted or produced in China, Russia, Iran or North Korea, subject to exceptions, yet domestic supply covers barely one percent of demand. Three developments in August 2026 reveal how the U.S. is racing to close that gap through a Lockheed-NioCorp scandium supply chain, a $104 million Utah heavy rare earth expansion backed by a conditional $725 million federal loan commitment, and a CSIS analysis documenting the real cost of dependency. None will be ready in time.
Introduction
On January 1, 2027, a procurement rule will take effect that could reshape the defense supply chain. DFARS 252.225-7052 bars defense contractors from delivering neodymium-iron-boron or samarium-cobalt magnets, or end items containing them, whose materials were mined, refined, separated, melted, or produced in China, Russia, Iran, or North Korea, subject to exceptions for most commercial off-the-shelf items, electronic devices, magnets made from recycled material in the United States, and official nonavailability determinations. The restriction reaches all the way back to the mine, and contractors must be able to support from their own records that no restricted step took place in a covered country.A magnet made from ore extracted in California but processed through a Chinese refinery does not qualify.
The gap between that requirement and current industrial reality is staggering. U.S. neodymium magnet demand ran at roughly 48,000 metric tons in 2025. Domestic supply that year was approximately 300 metric tons. Even with aggressive scaling by the two leading domestic producers, MP Materials and USA Rare Earth, announced U.S. capacity by end-2026 is only around 5,000 metric tons, about a tenth of total U.S. demand. The Center for Strategic and International Studies, in an April 2026 analysis, warned that without significantly more capacity within about eight months, meeting the January 2027 sourcing requirement may not be feasible.
Three recent developments illustrate both the urgency and the limits of the American response. On August 4, 2026, NioCorp Developments and Lockheed Martin, building on a Pentagon-funded development program with its Skunk Works division, signed a non-binding memorandum of understanding for scandium supply from a Nebraska mine that has not yet been built. On July 29, Energy Fuels began construction on a $104 million heavy rare earth expansion at its White Mesa Mill in Utah, the only conventional uranium mill still operating in the United States. And an April 2026 CSIS analysis quantified, in precise customs data, just how severely China's April 2025 export controls had already disrupted aerospace and semiconductor supply chains. Together, these three stories are not separate news items. They are facets of a single, escalating crisis in which the forcing function is the clock.
The Rule, the Gap, and the Yttrium Warning
The DFARS rule did not arrive without warning. Congress created the restriction in the FY2019 National Defense Authorization Act, extended it to mining, refining and separation in the FY2021 NDAA (Section 844), and then, in Section 854 of the FY2024 NDAA (Public Law 118-31), granted a one-year extension, shifting the expanded rule's start from January 2026 to January 2027, explicitly to give contractors more time to find compliant suppliers. The extension acknowledged what the Pentagon already knew: the domestic supply chain did not exist at the required scale. It has not materialized in the intervening year.
From January 2027 the rule's reach extends to mining, refining and separation, and its liability consequences are serious. Contractors who cannot support their provenance representations face contract termination risk and potential False Claims Act exposure. The Pentagon has described waivers as last-resort measures, though if the metallization pipeline remains offline at scale in January 2027, procurement officials may have no other instrument. The affected programs span a wide range of Pentagon weapons systems, from the F-35 to precision-guided munitions. The 2022 discovery of a Chinese-made samarium-cobalt alloy inside the F-35's turbomachine pump, which temporarily halted deliveries of America's most advanced fighter, showed the risks of complacency.
The CSIS analysis by Dr. Gracelin Baskaran and Meredith Schwartz goes beyond the magnet supply gap to document a more immediate casualty: yttrium. When Beijing placed seven categories of medium and heavy rare earths on its export restriction list in April 2025, including scandium, terbium, dysprosium, and yttrium, the effect on U.S. imports was nearly instantaneous. Chinese customs data cited by CSIS show that China exported just 17 metric tons of yttrium to the United States between April and December 2025, compared with 333 metric tons in the eight months before the controls took effect. A February 2026 uptick to 20 metric tons remained far below the more than 66 metric tons shipped in January 2025.
The downstream consequences were not abstract. Aerospace manufacturers use yttrium as a thermal barrier coating on jet engine components, protecting turbine blades from temperatures that would otherwise cause structural failure. Industry association AIA vice president Dak Hardwick stated directly that the supply chain's dependence on Chinese imports had contributed to rising costs and growing shortages. Consultant Kevin Michaels of AeroDynamic Advisory called it a "tangible example of how China is flexing its rare earth muscle." In semiconductors, where yttrium serves as a protective chamber coating, one industry figure rated severity at "9 out of 10." European prices for yttrium oxide rose about 4,400 percent between January and November 2025, according to Argus. China directly supplies 93 percent of U.S. yttrium consumption; the remaining 7 percent comes from material first processed in China. There is no meaningful domestic alternative.
The Skunk Works Bet: Scandium From Nebraska, If the Money Arrives
On August 4, 2026, NioCorp Developments announced that it had signed a non-binding memorandum of understanding with Lockheed Martin under which Lockheed could purchase up to 15 metric tons per year of scandium oxide or aluminum-scandium alloy from NioCorp's Elk Creek project in Nebraska over ten years. The agreement extends a $10 million Defense Production Act Title III award, made to NioCorp's subsidiary Elk Creek Resources Corp. under a joint development program with Skunk Works, designed to establish the United States' first domestic scandium mine-to-master-alloy supply chain. Tyler Robinson, vice president of technology roadmaps at Skunk Works, said the company looked forward to evaluating the supply as part of broader alloy development efforts.
Scandium's strategic importance is disproportionate to the quantities involved. Added to aluminum in fractions of a percent, it significantly improves strength, corrosion resistance, and weldability with almost no weight penalty, making it attractive for next-generation fighter airframes and additive manufacturing. Beijing placed scandium on its April 2025 export control list simultaneously with dysprosium, terbium, and yttrium. Global annual scandium output, even before those controls, was measured in tens of metric tons. The 15 metric tons contemplated in the NioCorp-Lockheed MOU would represent close to a fifth of 2024 global supply. NioCorp plans to produce approximately 100 metric tons per year from Elk Creek once fully operational, which would exceed total global 2024 output outright. As I noted in my coverage of the Pentagon's $400 million commitment to Australia's Sunrise Energy Metals in August 2026, the world currently has no commercial-scale primary scandium mine in operation anywhere, which is precisely why Lockheed hedged its position by also securing a parallel purchase option on Sunrise's Syerston project in New South Wales.
The Elk Creek project's commercial viability is not yet secured. NioCorp received a preliminary, non-binding indicative financing term sheet from the Export-Import Bank of the United States for up to $800 million in debt financing, and the company began review sessions with EXIM following publication of an updated feasibility study in August 2026. That study projects a $4.1 billion net present value over a 40-year mine life, with annual production including more than 8,000 tons of ferroniobium, approximately 118 tons of scandium, and more than 59,000 tons of titanium tetrachloride, alongside rare earth products. The mine portal construction program, budgeted at $44.6 million, began in February 2026, and the project is fully permitted for construction. But full production, the company estimates, is still 35 months from the start of full financing and construction. EXIM has designated Elk Creek a top priority and is reportedly operating at what NioCorp describes as "Trump speed," though no final commitment has been made. The MOU with Lockheed strengthens the demand narrative without eliminating the financing, construction, or production-ramp risks that separate a non-binding agreement from a functioning supply chain.
White Mesa and the Heavy Rare Earth Pinch Point
The more immediately consequential of the two midstream developments is Energy Fuels' expansion of its White Mesa Mill in Utah. On July 29, 2026, the company announced the commencement of construction on a $104 million Phase 1 circuit expansion designed to produce commercial-scale quantities of dysprosium and terbium, the two heavy rare earth elements that give high-performance neodymium magnets the thermal stability required in electric vehicle traction motors, wind turbines, jet engine actuators, missile guidance systems, and advanced radar platforms. Dysprosium and terbium production is targeted for completion by the end of 2027. Circuits for samarium, europium, and gadolinium are expected to follow by the end of 2028.
The strategic significance of White Mesa's expansion is difficult to overstate in the context of the Western supply chain. Most rare earth projects outside China are focused on neodymium and praseodymium, the light rare earths that constitute the bulk of permanent magnet material by weight. Heavy rare earth separation is considerably more technically demanding, and only a handful of facilities outside China, led by Lynas in Malaysia, have demonstrated it at production scale. White Mesa already holds commercial NdPr oxide separation capacity of up to 1,000 metric tons per year, and in December 2025 the company validated dysprosium oxide production meeting stringent magnet specifications, confirmed by a third-party end-user. CEO Ross Bhappu told Forbes that heavy rare earths are "sort of the holy grail" that allow magnets to maintain performance across temperature ranges, and identified them as the point where "China has an incredible grip on the market today." That grip extends not just to mining but to the metallization and separation infrastructure that took China decades to build and that Western competitors have only recently begun to replicate.
The expansion is underpinned by a conditional commitment for up to $725 million in senior-secured debt from the U.S. Office of Strategic Capital, announced June 18, 2026, with a 20-year loan tenor. Energy Fuels entered the second quarter of 2026 with working capital of approximately $960 million and has since announced a definitive agreement to acquire Vacuumschmelze, a German magnetics manufacturer, in a transaction valued at close to $1.9 billion in cash and stock. The acquisition of Australian Strategic Materials, which became legally effective in August 2026, adds metal and alloy capacity in South Korea, completing the vertical integration picture the company describes as the first true rare earth mine-to-magnet platform in the Western world. Rare earth oxides produced at White Mesa are expected to supply approximately 70 percent of the feedstock needed for Australian Strategic Materials' Korean operations.
The feedstock strategy for the Utah expansion relies substantially on monazite concentrate from the Donald Project joint venture in Australia, expected to deliver 8,500 to 9,500 metric tons per year beginning in 2028, subject to a final investment decision anticipated in the third quarter of 2026. The Phase 1 expansion is deliberately sized around that feedstock. A planned Phase 2 expansion, targeting 2029, would dramatically increase scale to 6,294 metric tons per year of NdPr, 80 metric tons of terbium, and 288 metric tons of dysprosium. The OSC's $725 million commitment is intended to fund both the White Mesa circuits and a rare earth metals and alloy facility to be constructed elsewhere in the United States. As I analyzed in my piece on the midstream processing bottleneck in August 2026, this pattern of government capital flowing to separation and conversion facilities, rather than mines, reflects a genuine policy consensus that the United States does not primarily have a mining problem; it has a processing problem.
One Deadline, Three Strategies, and the Math That Does Not Work
The DFARS deadline, the NioCorp-Lockheed MOU, and the White Mesa expansion share a single forcing function, but they operate on very different timelines. The deadline is January 1, 2027. The White Mesa heavy rare earth circuits targeting dysprosium and terbium will be complete, at best, by late 2027. NioCorp's Elk Creek mine, assuming EXIM financing is finalized and construction proceeds without delay, would reach full production approximately 35 months from financial close, which at current pace implies sometime in 2029 or 2030. MP Materials' much-anticipated 10X magnetics campus in Northlake, Texas, projected to produce 7,000 metric tons of finished magnets per year alongside 3,000 from its Fort Worth Independence facility, begins commissioning in 2028. The arithmetic is unambiguous: the January 2027 deadline cannot be met by domestic production at the scale defense procurement requires.
This is not a failure of ambition. It is a collision between the speed of capital deployment and the speed of industrial construction. The Pentagon is simultaneously the regulator imposing the compliance requirement and the investor funding the companies attempting to meet it. The $10 million DPA Title III award to NioCorp, the $725 million OSC commitment to Energy Fuels, the $550 million-plus partnership with MP Materials, and the $400 million conditional DoD loan to Sunrise Energy Metals in Australia all reflect the same "whole of government" logic: use federal capital to accelerate the private supply chain. But as the CSIS analysis observes, announcements are not qualified industrial supply. A term sheet is not a separation plant. A memorandum of understanding is not a metric ton of dysprosium oxide.
The likely outcome for January 2027, therefore, involves some combination of tightly controlled waivers for programs with no compliant alternative, higher prices for the scarce supply that does meet traceability requirements, and potential procurement delays for the most sensitive systems. A House bill introduced in May 2026 (H.R. 9073) would delay the expanded restriction to 2032, or earlier once DoD certifies enough non-covered suppliers exist, potentially providing additional runway, but that legislation is not enacted and its trajectory is uncertain. China's November 2025 suspension of the October export controls, reached as part of the Xi-Trump trade framework, provides temporary relief but no structural reassurance. CSIS is explicit on this point: the yttrium export collapse demonstrates that even when China formally suspends restrictions, it is not a reliable partner during periods of heightened geopolitical tension. The suspension runs through November 2026, which means it expires weeks before the DFARS rule takes full effect.
The convergence of heavy rare earth supply concerns across all three topics is not coincidental. Yttrium, dysprosium, and terbium are classified as heavy rare earths and scandium is usually grouped separately, but all four featured on China's April 2025 export control list, and all represent vulnerabilities far more acute than the better-publicized NdPr complex. Light rare earth supply, while still heavily dependent on Chinese processing, has at least one operating U.S. mine (Mountain Pass) and several international alternatives under development. Heavy rare earth supply, by contrast, has very little commercial-scale separation capacity outside China, led by Lynas in Malaysia, while U.S. circuits at White Mesa and MP Materials' Mountain Pass are still being built or commissioned. The entire Western defense industrial base is, in practical terms, betting that Energy Fuels finishes its terbium and dysprosium circuits before the next geopolitical inflection point.
Australia as the Indispensable Bridge
Running through all three developments is a common thread that often goes underappreciated in U.S.-focused coverage: Australia's role as the indispensable near-term feedstock source. Energy Fuels' White Mesa expansion is deliberately scaled around monazite from the Donald Project joint venture in Australia. Lockheed Martin holds a parallel purchase option on scandium from Sunrise Energy Metals' Syerston project in New South Wales, the same project that received a conditional $400 million Pentagon loan commitment in August 2026. CSIS describes Australia as the single most important U.S. partner in countering Chinese rare earth dominance, noting that the country attracted $64 million, approximately 45 percent of global rare earth exploration investment, in 2024, five times more than Brazil, and hosts 89 active projects compared with 18 in Canada and 12 in the United States.
This Australian dependency is a feature, not a bug, of current Western supply chain strategy, at least in the near term. Australian ore and concentrate can legally satisfy DFARS requirements provided the downstream processing occurs outside China, which is precisely the role White Mesa is being built to play. Monazite from the Donald Project feeds into Utah separation, which feeds into Australian Strategic Materials' Korean metal production, which could eventually feed into U.S. magnet manufacturing. The supply chain is genuinely multinational, traversing three continents, which creates its own logistical and geopolitical dependencies even as it reduces exposure to Beijing. The October 2025 U.S.-Australia Critical Minerals Framework formalized this partnership, and the string of Pentagon investments across both countries since then reflects a deliberate policy of treating the two nations as a single industrial base for critical minerals purposes.
Conclusion: The Gap Between Policy and Production
Recent months produced three genuine milestones in the American rare earth supply chain: a mine-to-warfighter MOU connecting Lockheed's Skunk Works to a Nebraska scandium deposit, a construction start on a U.S. facility designed to produce commercial-scale terbium and dysprosium, and a CSIS analysis documenting in precise customs data the consequences of failing to diversify sooner. Each represents a meaningful step. None closes the gap before the January 2027 DFARS deadline, and the CSIS data on yttrium serves as a live demonstration of what that gap costs when China decides to use its leverage.
The pattern across these developments confirms a structural reality that federal policy has been slow to internalize: building a domestic rare earth supply chain is not primarily a mining challenge. It is a processing, metallization, and vertical integration challenge that requires specialized infrastructure, long commissioning timelines, and patient capital at a scale the private sector cannot provide unassisted. The $725 million OSC loan to Energy Fuels, the $800 million EXIM application for NioCorp, and the broader $2 billion-plus White House critical minerals package are necessary responses, but they are responses measured in years, not months.
What January 2027 will actually look like in practice remains genuinely uncertain. Defense contractors will face a statutory obligation they cannot fully satisfy, an industrial base that is building but not yet built, and a waiver process that the statute treats as an exception but that logistics may require as a routine. The Pentagon's dual role as regulator and investor, simultaneously setting the compliance standard and funding the companies trying to meet it, reflects an honest acknowledgment that market forces alone will not close a gap this large on this timeline.
What is clear is that the rare earth crisis has moved decisively past its mining phase and into its processing phase. The three stories of August 2026 are ultimately one story: the United States is spending billions of dollars, on multiple continents, to build the midstream separation and metallization infrastructure that China spent decades developing and is now willing to restrict. The question is not whether that investment will eventually produce results. It almost certainly will. The question is whether the results will arrive before the deadline, or after the next geopolitical shock, or too late to matter for the programs that need them most.