A June 2 Council on Foreign Relations report finds that Project Vault, the Trump administration's landmark $12 billion critical minerals reserve, is insufficient on its own and harbours a structural paradox: in the near term, filling its shelves with processed materials would likely require purchasing from the very Chinese suppliers the programme is designed to render irrelevant. CFR authors Jonathan Hillman and Ishaan Thakker call for expanding the National Defense Stockpile, establishing a separate Strategic Resilience Reserve, and building plurilateral supply agreements through the FORGE framework.
Introduction
On a cold morning in February, General Motors CEO Mary Barra took her seat in the Roosevelt Room of the White House, flanked by executives from Boeing, Google, and GE Vernova, as President Donald Trump prepared to announce what his administration would bill as the most consequential strategic stockpiling initiative since the Korean War. The ceremony had the feel of a wartime mobilisation, and in some ways it was. Trump stepped to the podium and declared that American businesses had "risked running out of critical minerals during market disruptions" for too long. "Today," he said, "we're launching what will be known as Project Vault, to ensure that American businesses and workers are never harmed by any shortage."
The programme he unveiled was, by any measure, enormous. Project Vault would deploy twelve billion dollars, anchored by a fifteen-year, ten-billion-dollar direct loan from the Export-Import Bank of the United States, the largest financing deal in the agency's ninety-two-year history, more than double its previous record. Private commodity trading houses Hartree Partners, Mercuria, and Traxys would source the materials. Off-takers from aerospace to automotive would commit to long-term purchase agreements at specified inventory prices, essentially setting price floors that could make otherwise unviable mining and processing projects economically feasible. Boeing's executive vice president Jeff Shockey called it a step toward "supply chain resilience for critical minerals." EXIM Chair John Jovanovic described it as protection against "supply shocks" and a mechanism to "strengthen America's critical minerals sector."
Four months later, a team at the Council on Foreign Relations published a quieter document that asked the harder questions the Roosevelt Room ceremony left unanswered. Titled "The Stockpile Gap: How America Can Secure the Strategic Materials It Needs to Win," the June 2 report by CFR expert Jonathan Hillman and co-author Ishaan Thakker did not dispute the urgency that drove Project Vault into existence. It questioned whether the architecture being built was up to the task, and it identified a vulnerability at the programme's core that no amount of EXIM financing can easily resolve: in the near term, filling a strategic reserve with the processed materials American industry actually needs would almost certainly require buying from Chinese suppliers.
A Reserve Built on Borrowed Ground
To understand why the CFR's warning lands with such force, it helps to understand what Project Vault actually stockpiles, and what it does not. The reserve covers all sixty minerals on the 2025 US Geological Survey critical minerals list, with particular emphasis on rare earth elements, aluminium, antimony, copper, germanium, silver, and zirconium. These are not exotic laboratory curiosities; they are the physical substrate of modern defence and industrial technology, present in everything from the F-35's avionics to the batteries in a General Motors electric vehicle.
The problem, as Hillman and Thakker lay out with uncomfortable precision, is that the biggest chokepoints in critical mineral supply chains are not mines. They are refineries, processing facilities, and the specialised chemical operations that turn raw ore into the purified compounds that manufacturers can actually use. China controls up to ninety percent of the world's processing capacity across the critical minerals spectrum. For three specific types of rare earths necessary for heat-resistant magnets used in advanced weapons systems, China's share exceeds ninety-nine percent. The International Energy Agency estimates that China holds roughly a seventy percent market share for at least nineteen critical minerals and approximately ninety percent for cobalt, graphite, and lanthanide and yttrium rare earths.
This dominance was not accidental. It was built through decades of state subsidies, pricing practices that made it economically unviable for competitors to operate, and a strategic patience that Western commodity markets, oriented toward quarterly returns, were structurally unable to match. The result, as CFR President Michael Froman has described it, is "America's most dangerous dependence," one in which China "maintains significant economic leverage over access to inputs that are necessary for everything from everyday products like smartphones to advanced weapons systems like the F-35."
Project Vault's architects were not blind to this reality. The demand-driven structure, in which off-taker purchase commitments determine what is stockpiled rather than centralised government forecasting, was designed partly to ensure the reserve reflects actual industrial need. The price floor mechanism was designed to make allied and domestic projects viable by removing the volatility that has historically killed investment. But neither mechanism addresses what happens in the interval between today's supply landscape and the day when allied processing capacity can credibly substitute for Chinese output. Hillman and Thakker are direct on this point: meaningful processing independence "requires sustained investment at a scale not yet fully mobilized," implying a long-term effort rather than a short adjustment cycle.
The Warehouse Problem
Charles Boakye, an energy sustainability and transition analyst at Jefferies, frames the dilemma in terms that any manufacturer would recognise. "The bottleneck is not just the mining and sourcing; it's also the processing," he told Fortune. "Even if the U.S. can stockpile many of these materials, how and who is going to do the processing depending on the end-use applications?" Boakye called the stockpile a "first big step of many" needed to break China's stranglehold, but cautioned that the timeline for doing so spans at least three to seven years, at minimum.
The structural problem that Boakye identifies is not merely theoretical. In 2025, Ford was forced to halt production of its Explorer model due to a rare earths shortage, a vivid illustration of how supply disruptions now pose material risks not only to national defence but to the civilian manufacturing economy. The disruption came not from a mine closure, but from the processing chokepoint that sits between extraction and end use. A warehouse full of raw ore, absent midstream refining capacity, offers limited protection against precisely this kind of shock.
This is what the CFR report means when it observes that many existing stockpiles are "ore heavy," unable to directly support manufacturing without parallel investments in refining and magnet-making capacity. The gap between what can be quickly stockpiled and what manufacturers can quickly use is, in crisis scenarios, potentially decisive. CFR Senior Fellow Heidi Crebo-Rediker has put it starkly: current executive actions focus largely on expanding traditional mining and processing capacity, a necessary approach, "but one that takes years, often decades, and is insufficient to address potential escalation of tensions with China in the present."
The paradox embedded in Project Vault's near-term operation is therefore this: to fill its shelves with processed materials quickly enough to matter, the United States would likely need to source from Chinese processors, since the domestic and allied alternatives have not yet been built at scale. China's export controls on rare earths, which Hillman and Thakker describe as "a five-alarm siren that has not abated in Washington and capitals around the world," are precisely the vulnerability Project Vault was designed to hedge against. But if building the hedge requires transacting with the country whose controls created the vulnerability, the national security rationale becomes, at minimum, complicated. As I reported in my analysis of Project Vault's structural gaps in June, the question of what good a warehouse full of raw materials is if the factories needed to turn them into defence products do not exist applies with equal force to the processing question that precedes those factories.
Three Pillars: Expanding the Architecture
The CFR report's response to this predicament is not to abandon Project Vault but to situate it within a broader architecture that addresses the gaps its design leaves open. Hillman and Thakker advance four recommendations, three of which involve structural changes that would require congressional action and sustained executive commitment over years.
The first is expanding and modernising the National Defense Stockpile, the Department of Defense's existing strategic reserve of critical and strategic materials. The NDS was designed for wartime and national emergency scenarios; by statute, it is not intended to buffer the wider civilian economy or influence market prices. That narrow mandate was appropriate during the Cold War, when the United States possessed its own robust processing and refining infrastructure. It is increasingly mismatched with the current supply landscape. The NDS's decline since the end of the Cold War has been dramatic: its stockpiled inventory fell from 9.6 billion dollars in 1989, equivalent to roughly 21.9 billion dollars in today's money, to just 888 million dollars in 2021. Congress has taken some steps to reverse this erosion; the One Big Beautiful Act appropriated two billion dollars to the NDS Transaction Fund, and in 2025 the Department of Defense announced its intent to procure up to one billion dollars in stockpile materials. But the CFR's position is that these efforts, while welcome, remain insufficient to address the scale of the problem.
The second recommendation is establishing a Strategic Resilience Reserve, a separate commercial-facing stockpile designed to buffer the civilian economy from supply shocks that fall below the threshold of war or national emergency. The NDS, by design, cannot serve this function. Project Vault operates in the commercial space but with a structure and financing mechanism that limits its flexibility and scale. The SRS, as Hillman and Thakker envision it, would fill the gap between defence-specific reserves and the broader industrial economy. This proposal is directly aligned with the SECURE Minerals Act, which I examined in June in the context of the Congressional Research Service's sweeping resource guide and the legislative race to embed such a reserve into the National Defense Authorization Act; the convergence of CFR analysis and congressional momentum suggests the concept is gaining serious institutional traction.
The third recommendation, expanding private sector collaboration through Project Vault itself, is the least structurally novel but perhaps the most politically achievable. The programme's demand-driven design and price floor mechanism are, by Hillman and Thakker's account, genuinely innovative features that should be retained and scaled. The concern is not with the mechanism but with the assumption that it alone is sufficient.
The FORGE Dimension
The fourth and perhaps most consequential of the CFR's recommendations concerns the international architecture within which any American stockpiling effort must ultimately operate. Hillman and Thakker call explicitly for the United States to "explore plurilateral agreements to scale trusted sources of supply," language that maps directly onto the FORGE framework that Secretary of State Marco Rubio unveiled at the inaugural Critical Minerals Ministerial in Washington on February 4, 2026.
FORGE, the Forum on Resource Geostrategic Engagement, was conceived as a successor to the Biden-era Minerals Security Partnership, but with what the Trump administration has described as "sharper teeth and a commitment to speed." It brings together Australia, the European Union, Japan, South Korea, and fifty-one other nations in a plurilateral coalition designed to establish a preferential trade-and-investment zone for critical minerals, coordinated price floors, and a collective market presence large enough to offer a credible alternative to Chinese supply. As I reported in June, South Korea's chairmanship of FORGE expired at the end of that month, creating a governance inflection point for the alliance precisely when its coordinating function is most urgently needed.
Vice President Vance, speaking at the ministerial, articulated the underlying logic with unusual clarity: "Our goal here is to align trade policy, development finance, and diplomatic engagement towards a shared strategic objective. And that objective is very simple: diversifying global supply in the critical minerals market while strengthening the partner countries who help all of us in this shared effort." The CFR's alignment with this framing is explicit, and its analysts are equally candid about why the multilateral dimension cannot be bypassed. Even with strong multilateral support, it is nearly impossible, in their assessment, to "out-mine, out-process, and outspend China" acting alone. The bilateral frameworks the United States has concluded with Argentina, Morocco, Peru, the UAE, Australia, and Japan each contribute a piece; but it is the plurilateral architecture of FORGE that could eventually create the collective scale to shift processing economics in allied nations.
The Section 232 investigation launched by the Trump administration in April, examining whether to levy tariffs on processed critical mineral imports and their derivative products, adds a further dimension to this picture. A final decision is expected in mid-October. If tariffs are imposed broadly, they will accelerate the pressure on allied processing investment. If they are structured to exempt FORGE partners, they could become a powerful tool for directing processing capacity toward trusted-source nations. The outcome will shape the economic logic of allied investment in precisely the midstream capacity that Project Vault's near-term stockpiling problem requires.
The Long Game and Its Costs
Beneath all of these policy debates lies a temporal reality that no executive order or EXIM loan can compress. Building the processing capacity necessary to fill a strategic reserve with materials that did not flow through Chinese facilities takes years, in many cases decades. The permitting, capital formation, workforce development, and technological investment required to create a credible allied processing ecosystem represent a sustained national commitment of a kind that American industrial policy has not attempted since the postwar era.
This is precisely why analysts like Boakye are careful to distinguish between what Project Vault accomplishes now and what it could represent over time. "This is a first big step of many," he has said, adding that breaking China's stranglehold will take at minimum three to seven years. Crebo-Rediker's observation that executive actions focused on mining and processing expansion are "insufficient to address potential escalation of tensions with China in the present" is not a criticism of those actions; it is a statement about the fundamental mismatch between the pace at which geopolitical risk accumulates and the pace at which industrial capacity can be rebuilt.
The CFR report's framing of China's export controls as a "five-alarm siren" is carefully chosen. Sirens demand immediate response. But the response that the siren calls for, a fully diversified, allied-processing-anchored critical minerals supply chain, cannot be delivered immediately. What can be delivered immediately, and what Project Vault represents, is a holding action: a mechanism to buffer the civilian and defence industrial base against shocks while the longer investment cycle runs its course. The CFR's argument is not that the holding action is wrong. It is that the holding action is being treated as the strategy.
For its part, the Trump administration has shown little inclination to frame Project Vault as provisional. The Roosevelt Room ceremony, the EXIM record loan, the presence of Mary Barra and the Boeing executive team, all of it communicated permanence and scale. And indeed, the programme's most genuinely novel feature, the price floor mechanism that protects both investors and off-takers from the volatility that has historically made critical mineral projects economically unviable, is a structural contribution that deserves credit. As Boakye noted, the arrangement comes close to industrial policy of a depth that the United States has rarely been willing to attempt, which is precisely what makes it significant and what makes its limitations worth taking seriously.
Conclusion
In the weeks since the CFR report landed, Washington's critical minerals conversation has not slowed. The Section 232 investigation proceeds toward its October deadline. FORGE navigates its governance transition. The SECURE Minerals Act advances through the legislative machinery that the Congressional Research Service's June guide was designed to support. And somewhere in a government warehouse, the first materials procured under Project Vault's framework are beginning to accumulate.
The scene in the Roosevelt Room in February had a quality of resolution to it, as though the announcement of twelve billion dollars and a record EXIM loan represented a problem solved. Hillman and Thakker's report is a corrective to that impression, methodical and specific where the ceremony was declarative and broad. "Stockpiling is one of history's oldest responses to scarcity," they write, "but accumulating materials is no longer enough." True resilience, they argue, requires integrating reserves with production capacity, supply chain visibility, and technological upgrades, a system designed to identify critical bottlenecks, inform policy responses, and prevent economic coercion from escalating into crisis.
Boakye, the Jefferies analyst, offered a formulation that has stayed with me since I first encountered it. He called Project Vault a "first big step of many." It is a generous reading, and probably the right one. The danger is not that the step was taken; it was necessary and overdue. The danger is in mistaking a first step for a destination, in announcing a vault and forgetting to ask who holds the keys, and whether, for now at least, some of those keys are still in Beijing.