Analysis
Crusoe, the Denver-based AI data-center builder, closed a $3 billion-plus financing round at a $30 billion valuation, co-led by Atreides Management and Valor Equity Partners with participation from Mubadala Capital, according to Bloomberg and TechCrunch.
The round comes just ten months after Crusoe raised $1.38 billion at a $10 billion valuation last October -- tripling its valuation in under a year.
“The round comes just ten months after Crusoe raised $1.38 billion at a $10 billion valuation last October -- tripling its valuation in under a year.”
Founded in 2018 by Chase Lochmiller and Cully Cavness, Crusoe began by converting stranded natural gas from oilfield flaring into electricity for on-site data centers, then pivoted its infrastructure toward GPU cloud services as AI demand took off. The roughly 900-person company now counts Meta, Microsoft and OpenAI among its cloud customers, and a five-year, $13 billion contract to supply quantitative trading firm Jane Street with GPUs and AI infrastructure was a major factor pulling investors into this round.
Crusoe competes directly with Nscale, which is separately raising $3.5 billion toward its own IPO at an identical $30 billion valuation target, as well as CoreWeave, Lambda and Nebius -- the broader neocloud cohort racing to build GPU capacity faster than the hyperscalers can. Crusoe's edge is vertical integration: unlike pure GPU resellers, it also builds and operates its own power generation, letting it site data centers where energy is cheap rather than where fiber already runs.
The risk sitting underneath both this raise and Nscale's: neocloud valuations are being set by contracted revenue backlogs that assume multi-year AI compute demand holds at today's pace, and Nvidia has already trimmed one hyperscaler-adjacent commitment (its Ohio data-center guarantee for OpenAI) once this year after investor pushback. A $30 billion mark for a company that didn't exist as an AI infrastructure business three years ago prices in a lot of continued spend from a customer base -- Meta, Microsoft, OpenAI -- that all have their own reasons to eventually build more of this in-house.
