Crusoe Raises $1.375 Billion at $10B Valuation to Build the Clean-Energy Backbone of AI
AI infrastructure startup Crusoe has announced the initial closing of its anticipated $1.375 billion Series E led by Valor Equity Partners and Mubadala Capital, bringing its expected valuation to over $10 billion. The Denver-based company, which has divested its Bitcoin mining business, reports a power pipeline of over 45 gigawatts and counts Cursor, Fireworks, and Together AI among the customers of its cloud platform.
The AI infrastructure arms race has its newest billion-dollar entrant. Crusoe, a Denver-based company that began in crypto mining and now focuses on vertically integrated AI infrastructure, has announced the initial closing of its anticipated $1.375 billion Series E round, bringing its expected valuation to over $10 billion.
The round was co-led by Valor Equity Partners — the firm known for early bets on Tesla and SpaceX — and Mubadala Capital, the investment arm of Abu Dhabi’s sovereign wealth fund. Participants included NVIDIA, Founders Fund, Salesforce Ventures, Tiger Global Management, 137 Ventures, Altimeter Capital, and Franklin Templeton. The oversubscribed round is among the largest venture financings in the AI infrastructure space this year.
From Flared Gas to Frontier AI
Crusoe’s origin story is unlike most in an industry that tends to follow familiar playbooks. Co-founders Chase Lochmiller (CEO) and Cully Cavness launched the company in 2018 with a radically pragmatic idea: turn wasted natural gas from oil fields into energy for portable data centers used in cryptocurrency mining. The waste-to-compute model was environmentally smarter than drawing power from the grid and allowed the company to access stranded energy assets that utilities couldn’t profitably reach.
But the real transformation came as the AI infrastructure shortage hit critical mass. Crusoe divested its Bitcoin mining business entirely in 2024 and redirected the organization toward AI data centers and cloud services. The company has since fully exited its crypto operations to concentrate on building and operating data centers for AI workloads.
“Crusoe is in the business of activating energy for intelligence and helping the greatest innovators of our generation build the future faster,” said Lochmiller in announcing the Series E. The logic is identical to the mining days — secure cheap power at the source and build computing around it — but applied at a scale and for a purpose that dwarfs anything the company imagined in 2018.
A Power Pipeline That Rivals Small Utilities
The scale of Crusoe’s infrastructure ambitions is striking. The company’s forward power pipeline now exceeds 45 gigawatts, a fourfold increase from the prior year. To put that in perspective, 45 GW approximates the total installed generation capacity of several mid-sized European countries combined. It is a figure that would have seemed like science fiction for a startup just three years ago.
Crusoe’s flagship project is a 1.2 gigawatt data center campus in Abilene, Texas, whose first phase went live just one year after construction began. The company is also working on a recently announced 1.8 GW campus in Wyoming.
Beyond physical infrastructure, Crusoe operates Crusoe Cloud, a computing platform purpose-built for AI model training and high-throughput inference. The platform counts AI coding assistant Cursor, inference specialist Fireworks AI, and open-model platform Together AI as customers — all marquee names in the fast-growing layer of AI-native developer infrastructure. Crusoe Cloud bookings grew 5x in the first three quarters of 2025, the company said, a trajectory that helped make the oversubscribed Series E possible.
Why NVIDIA Wrote a Check
The presence of NVIDIA on the cap table deserves attention. Jensen Huang’s company rarely leads venture rounds, but it invests strategically in companies building infrastructure where its GPUs operate at scale. By backing Crusoe, NVIDIA is effectively betting that alternative AI cloud providers beyond the hyperscaler duopoly — AWS, Azure, and Google Cloud — will absorb a meaningful share of enterprise GPU demand.
That thesis has gained significant traction in 2026. Large enterprises and AI-native startups increasingly cite hyperscaler vendor lock-in and price opacity as reasons to evaluate specialized alternatives. Crusoe, TensorWave (which closed a $350 million round backed by AMD in June), CoreWeave, and Lambda Labs now represent a genuine cohort of challengers in what was, two years ago, effectively a monopoly market.
Mubadala Capital’s co-lead role is equally telling. The Abu Dhabi fund has emerged as one of the most aggressive Gulf sovereign wealth vehicles deploying into US AI infrastructure, consistent with the UAE’s national strategy to secure AI leadership through capital deployment. The sovereign backing gives Crusoe long-horizon patient capital that traditional venture funds struggle to provide for infrastructure projects with 5-to-10 year payback periods.
The Energy Question Defining the AI Era
Crusoe’s rise crystallizes what has quietly become the defining constraint of the AI buildout: not chips, not algorithms, not engineering talent, but power. Training frontier models and running inference at scale requires electricity at volumes that have overwhelmed grid planning cycles in Texas, Virginia, and across the Pacific Northwest.
The company’s energy-first model — sourcing power at the point of generation before it reaches the grid — addresses this constraint in ways hyperscalers cannot easily replicate. By building data centers adjacent to power plants rather than competing for existing grid capacity, Crusoe commissions facilities faster and at lower energy cost. The approach also reduces transmission losses and avoids the long interconnection queues that have bottlenecked competitors.
That said, owning large physical infrastructure carries risks that asset-light cloud businesses avoid. Crusoe’s Wyoming and Texas footprints expose it to regional grid reliability issues, extreme weather events, and water availability constraints for cooling — a resource that has already drawn regulatory attention in drought-prone areas of the American West. Unlike software companies, Crusoe cannot iterate its way out of a poor site-selection decision or a surprise shift in energy policy.
What the $10 Billion Valuation Signals
The $10 billion valuation places Crusoe in a category it did not occupy 18 months ago. The company is no longer a scrappy alternative cloud — it is a genuine infrastructure layer that enterprises and AI labs evaluate alongside hyperscalers in compute procurement decisions.
For the broader AI ecosystem, the Crusoe round sustains investor confidence in the physical infrastructure layer of the AI stack even as questions mount about returns on AI capital expenditure at the application layer. Infrastructure enabling AI workloads has proven a more defensible investment thesis than AI applications themselves, which face fierce competition, thin margins, and contested user relationships.
The next test is execution: whether Crusoe can operate at the 45-gigawatt pipeline scale it has committed to, while maintaining the cost structure and reliability standards that attracted customers like Cursor and Together AI in the first place. At $10 billion, the market has made its bet. The Abilene and Wyoming campuses will be the proving ground.
Correction (November 15, 2026): 9 sentences were corrected or removed to match the sources listed on this page.