On February 2, 2026, President Trump signed an executive order establishing Project Vault, a $12 billion public-private partnership backed by the largest single loan in EXIM Bank's 92-year history. The program targets all 60 minerals on the USGS 2025 Critical Minerals List specifically for civilian manufacturers. The Peterson Institute has since published a structural critique identifying three design flaws that could cause the program to fail precisely when it is needed most.
Introduction
On February 2, 2026, President Trump signed an executive order establishing Project Vault at a White House Oval Office ceremony attended by General Motors CEO Mary Barra, mining entrepreneur Robert Friedland, and senior administration officials. The Export-Import Bank of the United States simultaneously approved a $10 billion direct loan to anchor the program, the largest single transaction in the institution's 92-year history and more than double its previous record. Combined with approximately $1.67 billion in private capital from Hartree Partners, Traxys, and Mercuria Energy Americas, the total program size reaches $12 billion, making it the most aggressive U.S. strategic stockpiling initiative since the Korean War.
The program covers all 60 minerals on the USGS 2025 Critical Minerals List, an expanded roster that grew from 50 materials in 2022 to include copper, uranium, and lead in the current iteration. It is explicitly designed for civilian original equipment manufacturers: automotive companies, aerospace suppliers, energy technology producers, and electronics firms. This distinguishes it structurally from the existing National Defense Stockpile, which has been Pentagon-managed since 1988 and is accessible only during declared war or by order of the Undersecretary of Defense.
The Peterson Institute for International Economics published a policy brief approximately three weeks ago authored by Senior Fellow Cullen S. Hendrix that raised three structural concerns capable of undermining the program's core security rationale. Separately, the One Big Beautiful Act, signed into law on July 4, 2025, appropriated $2 billion to the National Defense Stockpile Transaction Fund, creating a parallel two-track architecture with its own set of coordination risks. The data now available on both tracks warrant a comprehensive assessment.
The Supply Shock That Prompted the Architecture
Project Vault did not emerge from a theoretical vulnerability. It emerged from documented, measurable supply disruptions that reached the American factory floor before any formal reserve existed. China's escalation of export controls across critical minerals followed a structured timeline beginning in 2023 with licensing requirements on gallium and germanium, extending to antimony restrictions in August 2024, and culminating in a December 3, 2024 directive that, for the first time, specifically prohibited exports of gallium, germanium, antimony, and superhard materials to the United States.
The real-world consequences were concrete. Ford temporarily ceased production of the Explorer at its Chicago assembly plant for one week in May 2025 due to rare earth magnet shortages caused by China's licensing system. Ford CEO Jim Farley described the situation as "day to day." In February 2026, the U.S. Air Force began receiving new F-35s without their next-generation AN/APG-85 radars due to a gallium sourcing delay attributable directly to Chinese export controls. These were not commodity price fluctuations; they were production stoppages at scale.
The quantitative dimension of the vulnerability is significant. China is the leading producer of 30 of the 60 minerals on the USGS 2025 list and maintains the leading refining position across 19 of 20 strategic minerals tracked by the International Energy Agency, with an average market share of approximately 70 percent. China controls more than 90 percent of global heavy rare earth processing. For gallium specifically, China accounts for approximately 99 percent of primary global output while operating at only three-quarters of estimated production capacity, meaning it retains the ability to increase supply pressure without expanding infrastructure.
China's antimony export restrictions provide the clearest before-and-after data point. Following the August 2024 restrictions, Chinese antimony exports fell by approximately 97 percent and global prices surged approximately 200 percent. Average monthly exports of unwrought gallium were 66 percent lower in the post-control period from January 2024 through October 2025 compared to the pre-restriction baseline. A Defense Logistics Agency assessment found net shortfalls across 88 materials valued at $14.83 billion, against a National Defense Stockpile that by 2023 held approximately $912 million in actual stockpiled materials covering just 6.2 percent of estimated wartime material shortfalls.
The November 2025 temporary suspension of China's export controls, negotiated following the Trump-Xi meeting in South Korea, paused the most aggressive restrictions until dates ranging from November 10 to November 27, 2026, depending on the specific control category. As I analyzed in "The November 10 Clock" in June 2026, that suspension is a tactical pause rather than a structural resolution: China's underlying export control architecture has grown more sophisticated with every passing quarter, and the ability to reactivate stricter controls remains fully intact.
How Project Vault Is Actually Structured
Project Vault's mechanics differ materially from the Strategic Petroleum Reserve and from the National Defense Stockpile in ways that create both operational advantages and structural vulnerabilities. The PIIE describes it accurately as "a hybrid of a futures market and a strategic reserve," and that hybrid character explains both its appeal to private investors and the concerns analysts have raised about its design.
Participating OEMs, which currently include Clarios, GE Vernova, Western Digital, Boeing, General Motors, Stellantis, and Google, identify which materials they require, at what grades and volumes, and commit financially to secure access during market disruptions. They pay upfront commitment fees covering their proportional share of storage costs and interest on the EXIM loan. Commodity trading firms Hartree Partners, Mercuria, and Traxys handle sourcing. Minerals are stored in secure, decentralized U.S. facilities. OEMs can draw down annual allocations for normal operational use, provided they replenish the reserve to original committed levels; in a major disruption, they may draw down their entire allocation, with full replenishment required afterward.
The program targets approximately a 60-day emergency supply buffer across covered materials. Release triggers are governed by predefined, transparent criteria agreed upon in advance. Senior administration officials told Bloomberg News the program was oversubscribed, citing investor confidence in participating manufacturers' credit quality, long-term purchase commitments, and the EXIM backstop. The ORF characterizes the OEM-driven financing model as shifting the fiscal burden from the state to the companies, noting that unlike traditional government stockpiles, Project Vault requires companies to fund their own supply security through commitment fees and carrying costs.
SCIS Director for Critical Minerals Security Gracelin Baskaran has noted that the structure aims to provide supply predictability while catalyzing domestic mining, processing, and refining without direct taxpayer subsidies. GE Vernova CEO Scott Strazik stated at the program's launch that "the need to grow supply chains and access to critical minerals has never been more important." The program is structured to allow broader allied engagement over time, and EXIM Chairman John Jovanovic participated in a Critical Minerals Ministerial hosted by Secretary of State Marco Rubio that was attended by representatives from more than 50 countries.
The PIIE Critique: Three Structural Vulnerabilities
Cullen Hendrix's May 2026 PIIE policy brief identifies three structural concerns that deserve serious analytical attention. Each connects to a different dimension of the program's design, and collectively they describe a scenario in which Project Vault could fail to provide the security buffer it promises at the exact moment of a genuine supply shock.
The first concern is voluntary participation. Hendrix argues that voluntary participation would exclude both large self-insuring firms and small enterprises unaware of their exposure, hollowing out the risk pool in ways that could cause the program to fail when it is needed most. Large manufacturers with diversified procurement strategies have limited incentive to pay commitment fees for coverage they can self-provide; small and mid-sized manufacturers may not understand their exposure to upstream critical mineral inputs. The result is a risk pool concentrated among mid-tier OEMs, which is neither comprehensive enough to distribute costs efficiently nor representative enough to mirror the actual structure of U.S. industrial dependence. PIIE recommends mandatory participation with fees scaled to firm size.
The second concern is the complexity of managing 60 highly differentiated minerals and their processed derivatives. As Hendrix states directly: "Storing 60 highly differentiated minerals and processed derivatives is far more complex than stockpiling oil because supplies can degrade over time and often need to be processed to be usable during a crisis." Rare earth elements exist across a spectrum of forms including oxides, metals, alloys, powders, and specialized magnets; these are not interchangeable commodities. A stockpile of rare earth oxides does not help a manufacturer who needs finished neodymium-iron-boron magnets for an electric motor. The gap between raw material storage and crisis-ready industrial supply is not a minor operational footnote; it is a central design challenge that the program's current structure does not fully resolve.
The third concern is the most structurally damaging to the program's stated rationale. Building reserves of processed materials in the near term would likely depend on Chinese suppliers, since China controls 90 percent or more of processing capacity for most of the minerals most critical to the stockpile's security purpose. This creates a circular dependency: the program designed to reduce U.S. exposure to Chinese supply disruptions must initially rely on Chinese supply to build the reserve that provides that protection. Hendrix frames the broader problem as a structural mismatch: "The United States is thus trying to solve a strategic infrastructure problem with tools designed for return-maximizing capital markets, which may create a structural mismatch between market expectations and US geopolitical objectives."
Additional concerns raised by analysts outside PIIE reinforce the critique. The Rare Earth Observer has described Project Vault as "a golf club, not a reserve, a system benefiting paying members rather than the broader economy," pointing to the member-access structure that limits broad economic benefit. Bloomberg Businessweek's June 1, 2026 issue warned that the planned stockpile "aims to ensure supplies of rare earths and other strategic elements, but it could upend markets instead," citing concerns about hoarding and counter-hoarding dynamics in thin markets where a visible government-backed buyer can drive price volatility rather than suppress it. The Le Billon and Deberdt research cited in Chemical and Engineering News notes that stockpiling can be "perceived as market manipulation, since governments rarely disclose stockpile volumes, which can fuel uncertainty and market instability."
The Two-Track Architecture: Project Vault and the National Defense Stockpile Expansion
The One Big Beautiful Act, signed into law on July 4, 2025, appropriated $2 billion to the National Defense Stockpile Transaction Fund for fiscal year 2025, expanding a reserve that had held approximately $912 million in actual stockpiled materials as recently as 2023. That $2 billion appropriation is the parallel track to Project Vault, and the two tracks are designed to serve different purposes: the NDS covers military and defense production needs, while Project Vault addresses civilian OEM supply security.
The broader OBBBA critical minerals package totals $7.5 billion across several instruments. This includes $5 billion to the Industrial Base Fund for critical mineral supply chain investments available through September 2029, $500 million for the Office of Strategic Capital's Defense Credit Program, and $1 billion for Defense Production Act financing through September 2027. The Pentagon separately announced intent to procure up to $1 billion in stockpile materials in 2025, including $500 million of cobalt, $245 million of antimony, $100 million of tantalum, and $45 million of scandium.
The two-track design raises a coordination problem that the Atlantic Council has identified precisely. If Project Vault and the expanded NDS are both competing for processed materials in markets where China controls the dominant processing share, they risk becoming competing buyers pursuing the same bottlenecked supply. The Atlantic Council has warned explicitly that Project Vault risks "becoming a second buyer, chasing the same bottlenecked materials that the defense stockpile is trying to" acquire. This is not a theoretical risk in thin markets for materials like processed heavy rare earths or antimony; it is a structural consequence of simultaneous demand expansion without commensurate supply-side investment.
The OBBBA's treatment of the Inflation Reduction Act's 45X tax credit adds a further complication. The legislation phases out the 10 percent production credit for applicable critical minerals beginning in 2031 and eliminating it in 2034. CSIS has noted that while OBBBA allocations represent meaningful supply-side support, the legislation does not deliver on the demand-side measures needed to drive sustained market demand. Carnegie Endowment analysts have argued that the OBBBA provisions fail to address the structural weaknesses in U.S. mining competitiveness, including permitting timelines, logistics infrastructure, and the economies of scale required to compete with Chinese processing.
As I analyzed in "The Stockpile Paradox" in May 2026, the structural trap confronting Western industrial policy is that both the United States and the European Union are racing to warehouse materials they can currently only source from the country they are trying to hedge against. The two-track architecture amplifies that trap by increasing aggregate demand from U.S. government-backed programs without a commensurate timeline for building independent processing capacity.
Governance Gaps, Oversight Risks, and the Credibility Problem
Project Vault currently operates without a statutory framework requiring participating companies to disclose reserve volumes. The opacity risk has been flagged specifically as a potential market manipulation concern by multiple researchers. The program relies on EXIM's existing authority rather than a new Congressional appropriation, and EXIM's mandate expires in December 2026, with reauthorization debates currently underway. If EXIM is not reauthorized or its mandate is narrowed, the institutional backstop underlying the $10 billion loan faces a discontinuity risk that would affect the program's governance architecture.
Historical precedent for this risk is documented. The Kennedy administration discovered in the early 1960s that the National Defense Stockpile contained approximately $7.7 billion in materials, roughly $3.4 billion more than estimated wartime requirements. The subsequent Symington hearings from 1962 to 1963 generated allegations of excessive profits and political favoritism. Congressional appropriators had used stockpile sales as budget relief, undermining the reserve's strategic function. The program that had been built as a national security instrument became a subject of political controversy precisely because its governance structure lacked transparency. Project Vault's accountability gap, specifically the absence of public disclosure requirements for stockpile volumes, destinations, and military-versus-civilian allocation, creates a similar vulnerability.
The credibility problem extends beyond domestic governance. PIIE's Hendrix has identified policy instability and concerns about U.S. coercive tactics as reasons why potential allied partners will be reluctant to anchor supply chain diversification strategies around U.S.-led initiatives. The Forum on Resource Geostrategic Engagement, announced at the February 2026 Critical Minerals Ministerial attended by representatives from over 50 nations, and the Pax Silica initiative launched in December 2025, represent the diplomatic architecture around these supply chain programs. But as Hendrix has written, "The difficulty facing US, and by extension, Western critical minerals strategy is not primarily one of resources or capital but of credibility." A program that may need to source its initial inventory from Chinese suppliers, operates without full statutory authority, and faces EXIM reauthorization uncertainty by December 2026 has a credibility deficit that diplomatic frameworks cannot easily compensate for.
The Council on Foreign Relations' Jonathan Hillman, writing on June 2, 2026, recommended that the United States expand and modernize the NDS, establish a separate strategic resilience reserve for commercial needs, expand Project Vault private sector collaboration, and explore plurilateral agreements. That recommendation implicitly acknowledges that the current two-track structure requires further architectural refinement to function as an integrated system rather than two parallel programs pursuing overlapping objectives with insufficient coordination.
Forward Outlook: What the Data Require
The demand environment that Project Vault is designed to address is growing structurally, not cyclically. Global consumption of magnet rare earth oxides is projected to quadruple from $9.6 billion in 2025 to $44.1 billion by 2040. FGS Global research finds that 87 percent of engaged Americans are concerned about Chinese mineral market control and 96 percent consider U.S. access to critical minerals important. The underlying demand is real, the supply concentration is documented, and the disruptions have already produced measurable production stoppages.
The structural question the data raise is whether a 60-day buffer stockpile of 60 differentiated minerals, built partly through Chinese suppliers, governed by voluntary OEM participation, and anchored to an EXIM loan that expires in December 2026, can provide durable supply security against a counterparty whose export control architecture has grown more sophisticated every quarter since 2023. The Atlantic Council's assessment is precise: "The administration delivered speed and ambition. Execution and durability are the harder tests ahead."
The PIIE's three recommendations, mandatory participation with fees scaled to firm size, funding calibrated against worst-case collective supply shocks rather than individual firm shocks, and prioritization of processed materials over raw ore, describe the modifications required to close the most significant design gaps. The Atlantic Council's warning about the dual-buyer problem between Project Vault and the NDS describes the coordination requirement. The EXIM reauthorization timeline and the absence of a statutory disclosure framework describe the governance requirements.
The $12 billion headline figure and the EXIM record are significant. The participation of Boeing, GM, GE Vernova, and Western Digital provides genuine demand-side credibility. The oversubscription signal from private investors confirms that the market structure is sound enough to attract capital. But the PIIE critique is supported by the data, not contradicted by it: a program designed to reduce dependence on Chinese-controlled processing capacity cannot achieve that objective if its near-term inventory must be sourced from that same capacity. That is not a failure of ambition; it is a consequence of the timeline mismatch between the urgency of the stockpile need and the years required to build independent processing infrastructure at scale. The 60-day buffer, correctly calibrated and governed, buys time. It does not resolve the structural dependence it was designed to hedge.