Crusoe building turbine-powered AI campuses

Crusoe is moving from finding cheap power where it exists to building AI campuses around power it directly controls. That changes the bottleneck from waiting on utility interconnection queues to ordering turbines, fuel, and site infrastructure up front. The GE Vernova and PROENERGY fleets together give Crusoe a preplanned path to stand up very large campuses with hundreds of megawatts at a time, instead of piecing together capacity from the grid.1 2 3 4 5

• The GE package is not a small backup system. It combines 29 LM2500XPRESS aeroderivative gas turbines, 10 ordered in December 2024 and 19 more booked in June 2025, for nearly 1 GW. These are fast deployable units meant to run as primary on site generation for data center loads.1 2
• The PROENERGY deal adds another 13 PE6000 turbines at 50 MW each, or 650 MW, with deliveries by summer 2027 for hyperscale projects. In practice, Crusoe is assembling a fleet of modular power blocks that can be dropped next to new compute clusters as demand lands.3 4
• This starts to make Crusoe look more like the largest AI infrastructure builders, where power procurement is the product. CoreWeave disclosed about 3.1 GW of contracted power capacity as of December 31, 2025, showing that winning big GPU contracts increasingly depends on locking power before servers arrive.5 6

The next phase is a more vertically integrated AI factory model, where the winning providers are the ones that can line up turbines, fuel, storage, cooling, and GPUs as one deployment package. For Crusoe, that should expand its edge from lower energy cost into faster campus delivery, which is becoming just as important as price.1 3 5 6