Analysis
Crusoe has abandoned a $1.25 billion plan to buy gas turbines from Boom Supersonic for its AI data centers, according to TechCrunch, with Boom's own CEO Blake Scholl confirming the collapse on X.
Pulse has tracked Crusoe's financing and infrastructure buildout closely since its $3 billion Series F at a $30 billion valuation -- what's changed since that coverage is the company's actual power-sourcing strategy for the data centers that funding round was raised to build.
Crusoe had agreed to buy 29 of Boom's 42-megawatt Superpower turbines -- machines that share about 80% of their parts with Symphony, the engine Boom is developing for its Overture supersonic jet -- with first deliveries slated for 2027. Scholl says turbines are simply "no longer part of Crusoe's near-term primary power mix" at its Abilene, Texas campus; Crusoe confirmed it is no longer doing business with Boom at all, citing flexibility across wind, solar, batteries and the grid.
What This Says About The AI Power Crunch
Crusoe, founded in 2018 as a bitcoin miner running on stranded natural gas, has spent the past two years becoming one of the largest AI data-center builders -- the same company Value Add VC covered alongside Anthropic's $15 billion debt facility as part of the broader neocloud financing wave. Its walkaway from Boom isn't a sign the power crunch eased; if anything it's the opposite -- a sign operators are still actively shopping across gas turbines, grid interconnects, batteries and renewables rather than locking into any single novel technology this early.
For Boom, the loss stings more: the Superpower turbine business exists specifically to commercialize engine technology built for Overture, and losing a $1.25 billion anchor customer months after raising $300 million to fund that pivot is a real setback, even though Scholl says other customers remain in the pipeline and the company still targets 250 megawatts delivered next year, scaling to 1 gigawatt by 2028.
What the cancellation doesn't tell us yet: which power source Crusoe is actually substituting in Boom's place, or whether this reflects turbine-specific problems -- cost, timeline, technology risk -- versus a broader rebalancing away from any single vendor. Either read is a caution for any startup selling novel power hardware into the AI buildout: neocloud operators are proving willing to walk from nine-figure commitments the moment their own math changes.
