Crusoe, the Denver-based AI cloud and data center company, has raised more than $3 billion in a new funding round that values the startup at roughly $30 billion, Bloomberg reported on September 3, 2026 (UTC-4). The valuation has tripled in less than ten months — a pace that tells you everything about how desperate the AI industry is for compute capacity.
From $10B to $30B in ten months
In October 2025 (UTC), Crusoe closed a $1.38 billion Series E round at a valuation of roughly $10 billion. That round was co-led by Mubadala Capital and Valor Equity Partners, with NVIDIA, Fidelity, and Founders Fund participating.
The new round is co-led by Atreides Management and Valor Equity Partners, with Mubadala Capital — the investment arm of Abu Dhabi's sovereign wealth fund — returning as a participant. The round has been finalized, according to people familiar with the situation cited by Bloomberg.
Crusoe's total funding now exceeds $4.4 billion across its lifetime.
Who is Crusoe, and why does everyone want a piece?
Crusoe is not a model company. It does not build chatbots or release benchmark scores. It builds and operates AI-optimized data centers, with a twist: the company pursues an "energy-first" strategy, locating facilities near stranded or underutilized power sources — flare gas at oil wells, hydroelectric dams, wind and solar farms — and converting that energy into compute.
The name comes from Robinson Crusoe, the castaway who made the most of limited resources on a deserted island. The analogy is deliberate: Crusoe treats energy as the scarce input and builds compute around it, rather than the other way around.
Its customer list reads like the who's who of AI: OpenAI, Microsoft, and Meta all run workloads on Crusoe infrastructure. Crusoe is also one of the primary builders and operators of Stargate, OpenAI and Oracle's $500 billion-plus initiative to construct next-generation AI compute capacity across the United States.
The Jane Street deal
Weeks before this funding round closed, Crusoe signed a five-year, $13 billion cloud agreement with Jane Street, the quantitative trading firm. The contract supplies Jane Street with GPU clusters and high-performance computing capacity — a signal that AI infrastructure demand is no longer coming only from model labs. Financial firms, biotech companies, and enterprise customers are now competing for the same GPU pools.
That deal alone is larger than the entire annual revenue of many public cloud providers' AI divisions.
What the money is for
The new capital will fund Crusoe's aggressive expansion. The company is building gigawatt-scale data center campuses, including a major facility in Abilene, Texas, developed through a $15 billion joint venture with Blue Owl Capital and Primary Digital Infrastructure. It has also expanded into Europe with its first data center in Norway, partnering with Polar Energy.
Crusoe has also introduced "Crusoe Spark," modular AI data centers designed for edge computing deployments — smaller, portable units that can be placed near energy sources without requiring full-scale campus construction.
Why this matters
The AI infrastructure gold rush has moved beyond model labs. Companies that own the physical layer — data centers, power connections, cooling systems, and the operational expertise to keep GPU clusters running at full utilization — are now capturing enormous valuations. Crusoe's tripling in ten months reflects a simple reality: there is not enough compute to go around, and whoever can deliver it reliably and cheaply holds the leverage.
For model companies like OpenAI and Anthropic, infrastructure costs are the single largest expense. For investors, Crusoe represents a way to bet on the AI boom without picking which model wins — every major lab needs compute, and Crusoe sells to all of them.
The risk is that the buildout outruns demand. Crusoe and its competitors are committing tens of billions to new capacity on the assumption that AI workloads will keep growing exponentially. If that growth slows, the data center sector could face the same overcapacity cycle that hit telecom in the early 2000s. For now, though, the market is pricing in scarcity, not glut.
No comments yet