Crusoe drops $1.25B plan to power AI data centers with Boom turbines

Crusoe, a developer of AI data centers, has walked away from a plan valued at roughly $1.25 billion to deploy Boom Supersonic turbines as power sources for its facilities, according to a TechCrunch report published September 25. The disclosure came from Boom Supersonic CEO Blake Scholl, who said the company's new stationary power plants are no longer part of Crusoe's near-term plans.

The report offers no stated reason for the change and does not name a replacement supplier. What is clear is the scope: a billion-dollar-plus power arrangement that Crusoe had been pursuing will not move forward in the near term, leaving the data center operator to find electricity elsewhere for the capacity involved.

The shift lands against a backdrop of soaring electricity demand from AI. Larger training runs and always-on inference workloads have pushed single campuses toward power draws that strain local grids, making generation capacity an increasingly hard constraint on how fast compute can be built. Operators have been weighing gas turbines, fuel cells, nuclear and other routes, and Crusoe's decision shows those paths do not all land as planned.

Boom Supersonic is best known for supersonic passenger aircraft; stationary power generation is a newer extension of its business. Losing a near-term customer role is a setback for proving commercial traction beyond aviation. The TechCrunch framing is specific: Boom's plants left Crusoe's near-term plans, not the industry's need for gas-fired generation as a category.

For Crusoe, shelving a billion-dollar power plan usually reflects a fresh trade-off among construction timelines, grid interconnection and cost structure. Because the report does not detail those factors, the safe reading is that the plan is on hold — not that gas-fired power has been ruled out, and not that a different vendor has been selected.

The next signal to watch is substitution: whether Crusoe turns to another turbine maker, builds its own generation, or signs long-term power purchase agreements. Each path implies a different schedule for bringing capacity online, and for Boom the question is whether stationary power can convert interest into firm orders.

Power now sits alongside chips and capital as a gating factor for AI infrastructure. The Crusoe-Boom split is a reminder that the next phase of the compute race may be decided by who solves electricity most reliably and cheaply.