The Columbia-class ballistic missile submarine is the next generation of American nuclear deterrence. Twelve of these boats will replace the aging Ohio-class fleet, entering service over the 2030s and 2040s, each carrying 16 Trident IIs and driven by a ghost-quiet electric motor that renders them acoustically invisible to any adversary. What makes all of that possible — the propulsion, the stealth, the strike precision — depends almost entirely on rare earths refined in China. This is perhaps the Navy’s most consequential and least discussed vulnerability.
The dependency runs through every layer of the capability stack. The sub’s permanent magnet motor requires significant quantities of dysprosium and neodymium. Neodymium is the base element in the world’s most powerful permanent magnets. Dysprosium is what keeps those magnets functional under heat and stress. Without it, the magnets demagnetize under the sustained temperatures of operational use, and the motor fails. In addition, the acoustic systems that will give the Columbia-class its stealth advantage — sonar transducers, vibration dampeners, noise isolation arrays — rely on terbium and dysprosium-based magnetostrictive materials. The guidance systems in the submarine-launched missiles they carry depend on samarium-cobalt permanent magnets. China controls the refining capacity for all of it.
This is a supply chain dependency that is far less addressed than the radar and semiconductor vulnerabilities that have consumed Washington’s attention, with materials such as gallium nitride. To solve the U.S. Navy’s subsea rare earth vulnerability, the United States needs a financial architecture of offtake commitments, loan guarantees, and milestone-based co-investment that restores the basic market conditions that four decades of Western offshoring and Chinese industrial investment have hollowed out.
To be upfront, I invest in seed-to-Series B defense tech startups and have spent much of the past few years focused on maritime and critical materials. I hold active positions in startups that could benefit directly from the policy outcomes I recommend in this piece. Readers should weigh that overlap. However, my argument is not for any specific company to receive a contract, but for the capital architecture that removes the market failure preventing the best companies — including ones I have no stake in — from being built at all.
Beijing’s Card
Transforming mined ore into the high-purity rare earth metals that go into permanent magnets is not a single step. It is a multi-stage chemical gauntlet: Ore containing as little as two or three percent rare earth content should be concentrated, then subjected to thousands of sequential stages of solvent extraction to separate individual elements. Since rare earth elements are chemically nearly identical to one another, requiring precise, iterative chemistry using toxic acids and solvents that generate significant hazardous waste, this process is extremely difficult. For every ton of rare earth oxides refined, the process generates approximately 2,000 tons of toxic waste. The result is that as of 2024, China controlled 91 percent of global rare earth separation and refining, and approximately 99 percent of heavy rare earth processing specifically.
China didn’t come to dominate this process through superior geology. It got there first through structural macroeconomic advantages — state-directed capital, currency policy, and infrastructure investment at an irreplicable scale. On top of that came roughly $10 billion in direct Chinese government subsidies between 2010 and 2019 alone, combined with subsidized inputs and a tolerance for environmental costs not shared in the West. The economics of rare earth processing, as currently done, do not work without state support. Since China holds a near-total monopoly on heavy rare earth processing, it can exploit this structural imbalance assertively.
For instance, around 2010, China began restricting rare earth exports (triggered by a diplomatic spat with Japan), sending prices soaring. Dysprosium oxide climbed from roughly $100 per kilogram to above $2,300. That spike attracted a Western investment response. Molycorp raised hundreds of millions of dollars in capital and took on additional debt, restarted the Mountain Pass rare earth mine in California, and began building out processing capacity. Though the Mountain Pass is primarily a light rare earth deposit, rich in neodymium and cerium but not in the heavy rare earths — dysprosium, terbium — that the defense supply chain needs most, market failures persisted. Once the new mine’s capacity approached operational scale, China flooded the market with cheap rare earth oxides at prices that unsubsidized Western producers couldn’t match. Dysprosium fell back below $200 per kilogram, and Molycorp, unable to service its debt at those prices, filed for bankruptcy in 2015. The pattern of letting prices rise until a Western competitor commits capital, then lowering prices until that competitor fails, is not unique to rare earths. In fact, nickel and lithium tell similar stories. Whether this constitutes deliberate strategic manipulation or simply the predictable consequence of state-directed overproduction operating without profit constraints is a debate worth having. The effect on Western industrial capacity has been the same either way.
While companies like MP Materials and Lynas have begun rebuilding light rare earth processing capacity in the West, no comparable alternative exists for the heavy rare earth midstream that the Navy’s propulsion and stealth stack depends on. The Columbia-class will carry more permanent magnet content per hull than any predecessor, expanding rare earth demand further. But without these minerals, the sub can’t move or stay silent. It won’t ever have the chance to shoot.
Since December 2024, Beijing has imposed successive export restrictions on critical minerals. While some controls have been partially suspended under trade negotiations, exports to U.S. military end-users have remained banned. Beijing’s June 2026 decision to place MP Materials and USA Rare Earth — the two most prominent American companies building domestic rare earth processing capacity — on its export control list makes clear that China is actively targeting the companies trying to solve the problem.
The Capital Gap
The U.S. government has begun to respond. Defense Production Act Title III has funded extraction and processing grants, including a $29.9 million award to ElementUSA Minerals. Project Vault, announced in February 2026, commits $12 billion in public-private capital to strategic mineral stockpiling, the largest U.S. minerals initiative in a generation.
These are meaningful steps, but it is worth being precise about what they are and are not. A stockpile of processed materials buys strategic buffer time, not domestic industrial capacity. The prescription in this piece is different in kind. Offtake commitments, loan guarantees, and binding procurement floor prices provide a financial architecture that makes it economical to build and operate domestic processing facilities. The case for a broader federal processing initiative — including permitting reform and workforce investment — has been made persuasively in these pages. What is missing is the specific financial architecture for activating private capital.
I evaluate companies in the critical minerals processing and naval systems supply chain regularly and consistently encounter a cohort of technologically credible companies — novel magnet chemistries, rare earth recycling processes, alternative magnetostrictive material development — that are early-stage, undercapitalized, and structurally unable to attract the institutional investment they need. In most cases, the Department of Defense demand signal is too weak or inconsistent to de-risk. A venture fund will not lead a seed round in a rare earth processing startup if the primary customer cannot articulate what, how much, and when it will buy that good.
The Replicator program is surprisingly instructive for rare earths. The program faced its fair share of troubles, including congressional investigations documenting execution failures, immature selections of systems, software problems, and fielding rates below stated goals. But Replicator at least told the market, with enough specificity and dollars behind it, that the Department of Defense would buy autonomous attritable aerial systems at scale. As a result, more than 500 commercial companies competed for Replicator-related contracts, thirty received awards, and capital certainly followed the demand signal.
No such signal exists for rare earth magnet supply chains or the processing capacity they require. While investors and founders know the strategic case for such capacity, they do not believe the government will actually buy, consistently, at the prices and volumes required to justify building.
Building a Market Doesn’t Mean Picking Winners
The recently established Economic Defense Unit carries a relevant mandate of aligning defense strategy with economic competition and securing access to critical capabilities. The Economic Defense Unit should own the rare earth propulsion dependency problem, functioning as a demand signal generator capable of making multi-year, binding offtake commitments.
The Office of Strategic Capital was created to attract private investment to national security sectors, and it also has the tools to structure the debt side of this equation. Loan guarantees for rare earth processing facilities — similar in structure to the Department of Energy loan guarantee program that catalyzed the early cleantech wave — would lower the cost of capital for companies building domestic or allied processing capacity. The Office of Strategic Capital can provide both growth-stage and early-stage investors with those stage-appropriate market anchors if it prioritizes the rare earth supply chain with the same urgency currently applied to semiconductor manufacturing. The June 2026 conditional loan commitment to rare earth separation and metallization startup Phoenix Tailings — $500 million from the Office of Strategic Capital, structured to catalyze an additional $500 million in private capital — signals it is becoming a priority.
None of this is simple to execute. The Economic Defense Unit was only officially launched in April 2026 (though it had been operating informally before that date), and its mandate is still being operationalized. The relationship between the Economic Defense Unit and the Office of Strategic Capital remains to be defined. Both offices have tools applicable to critical minerals investment, but interagency coordination on defense industrial base priorities has historically been weak. Multi-year offtake commitments and loan guarantees also require appropriations authority and on Congress that aren’t yet present.
Urgency is warranted, but its form matters enormously. When facing China’s centrally planned dominance across a number of markets, some are tempted to copy it and enable the government to take equity stakes in rare earth champions, propping them up subsidy-for-subsidy. The Trump administration gestured in this direction with its equity stake in Intel. I think that’s the wrong model because concentrating nearly $11 billion in a single incumbent doesn’t fix the underlying market structure. The rare earth supply chain also doesn’t have a single-champion solution. It requires building viable markets for dozens of companies across mining, processing, and manufacturing, and so equity stakes likely won’t fix it. The correct fix is neutralizing the specific weapon China keeps using: the ability to crash prices the moment a Western competitor approaches operational scale. Rather than picking winners, offtake agreements, loan guarantees, and binding multi-year procurement commitments restore the basic market conditions that enable competition.
At the endlessly innovative startup layer, the defense ecosystem has developed reasonably functional channels for software and sensor companies to signal demand. For example, Small Business Innovation Research grant Phase II, Defense Innovation Unit other transaction authority contracts, and the Air Force’s AFWERX program. But for hardware-intensive critical minerals companies, those mechanisms are a poor fit. A rare earth processing startup with a novel extraction chemistry and $2 million in revenue is not scalable through other transaction authority contracts alone, unattractive to most institutional investors without an anchor customer, and not large enough to navigate traditional defense acquisition. Defense Production Act Title III authority exists to address exactly this gap, but it has been applied too narrowly by focusing on downstream production rather than on the midstream processing layer. A purpose-built Defense Production Act Title III program targeting rare earth magnet manufacturing, combined with milestone-based equity co-investment from the Economic Defense Unit, would create effective private capital conditions.
The allied dimension deserves substantial attention, too. Australia, Japan, Canada, and the United Kingdom possess either rare earth deposits, nascent processing infrastructure, or both, but diplomacy often lacks a commercial financing mechanism. The Development Finance Corporation is the right vehicle, but it is not being used at scale. Despite a 2025 reauthorization that expanded its investment capacity to include financing projects in high-income allied nations, and diplomatic frameworks now in place with both Japan and Australia on critical minerals cooperation, recent Development Finance Corporation rare earth commitments have flowed to Brazil and Angola rather than to the allied industrial partners best positioned to scale midstream processing.
Conclusion
The technology and companies to break America’s rare earth processing dependency on China — and with it, China’s ability to hold the Columbia-class submarine program at risk — exist. What is missing is the sustained, structured demand signal that tells private capital this market is real and so too is the financial architecture.
Defense Production Act Title III grants, Office of Strategic Capital loan guarantees, Development Finance Corporation co-financing with allied partners, and multi-year offtake commitments from the Defense Department can together create the market conditions private capital needs. None of that requires copying Beijing’s playbook. It requires treating rare earth processing with the same urgency Washington currently reserves for semiconductors. The first Columbia-class submarine is scheduled to go on patrol in 2030. Building rare earth processing capacity from commitment to operational scale takes five to seven years. A decision made today arrives barely in time. A decision deferred by two years does not. Getting that architecture right is the precondition for building the Navy that the United States will need, and for doing so before Beijing reminds the United States of what it forgot to build.
Ryan Musto is an investor at Alumni Ventures, where he helps lead investments in DeepTech and DefenseTech with a focus on maritime systems, critical materials, and advanced manufacturing. He holds an M.Sc. with Distinction from the University of Oxford and a B.A. magna cum laude from Cornell University.
Image: Lt.j.g. Richard Locklear via DVIDS.
