U.S. Department of Defense Plans $5 Billion Loan to AI Cloud Computing Startup Fluidstack
nashnova research
The U.S. Department of Defense is negotiating a roughly $5 billion loan to AI cloud startup Fluidstack to bolster domestic data-center supply chains — if finalized, it would be the largest single loan ever issued by the Pentagon's Office of Strategic Capital, signaling that military funding is now a live channel for AI infrastructure finance.
What is the money for — not building data centers?
Fluidstack plans to spend the funds on U.S. supply-chain and manufacturing capacity for data-center components, not on constructing new AI facilities directly.
This means → the bottleneck isn't the buildings — it's what goes inside them: cooling, power equipment, networking hardware, and their domestic production.
The loan application is being advised by Erebor, a new national bank founded by Oculus creator Palmer Luckey, focused on tech and defense clients.
Where does the Pentagon's money come from — what is the OSC?
The Office of Strategic Capital was established in 2022 under the Biden administration to finance companies with national-security applications through loans and loan guarantees.
The Trump administration and the One Big Beautiful Bill dramatically expanded its mandate: authorization to deploy over $200 billion in financing over the coming years, plus permission to hire Wall Street veterans at competitive pay.
In plain terms = this isn't traditional defense procurement. The Pentagon effectively opened a "national-security bank" — lending directly to companies, sometimes taking warrants or equity stakes that give the government an ownership interest.
Who is Fluidstack — how does a London startup get a Pentagon loan?
Fluidstack was founded in London in 2017, pivoted last year to focus on U.S. AI infrastructure, relocated its headquarters to New York, and now employs roughly 700 people — all based in the U.S.
The company has signed multi-year data-center site leases in New York, Indiana, Louisiana, and Texas.
This reflects a broader pattern: in the AI infrastructure race, companies are fast-tracking their U.S. footprint to unlock government funding and top-tier customer relationships.
Who is backing Fluidstack — what does the investor lineup signal?
Earlier this year Fluidstack closed an $830 million round led by Leopold Aschenbrenner's Situational Awareness fund.
More recently it completed a $1.5 billion round led by Jane Street, at a valuation of $18 billion.
Core demand comes from broad partnerships with Google and Anthropic — Google provides access to TPUs (Tensor Processing Units — Google's custom AI chips) and backstops Fluidstack's data-center leases, helping to lower financing costs.
This means → Fluidstack is not an empty shell chasing government money: long-term contracts with top-tier AI companies are the credit anchor that makes a Pentagon loan viable.
What does this mean for the market — the $4.1 trillion big picture?
JPMorgan analysts estimate that by 2030, lenders will provide a combined ~$4.1 trillion in financing for AI infrastructure projects.
Military capital entering the space marks a new expansion of AI infrastructure funding — from tech giants' own balance sheets + venture capital → now extending into the defense-budget system.
The deal is not yet finalized; terms including the interest rate may still change. In plain terms = the direction is clear — the U.S. has elevated AI compute to the level of a "strategic national asset," and the Pentagon is willing to write the check; the exact number on that check is still being negotiated.
市场有风险，内容仅供研究参考，不构成投资建议。