Crusoe, a startup helping build the first data center for Stargate, says it secured a deal with a gas turbine owner to generate 4.5 gigawatts of energy by 2027
Anissa Gardizy / The Information :
Context & Ripple Effects
Crusoe’s earlier business used excess gas from oil extraction for crypto-mining data centers, as covered in its $505M Series C financing for that gas-powered model. This agreement moves the company’s power strategy from dispersed, opportunistic generation toward capacity sized for a major AI campus.
The arrangement matters because Crusoe is participating in Stargate’s first facility: securing generation is not ancillary to the buildout, but a prerequisite for turning data-center construction into usable compute capacity.
First-order effects
- Crusoe gains a stated route to 4.5 GW of generation by 2027, giving its Stargate-related buildout a defined power-supply plan rather than relying solely on grid availability.
- The gas-turbine owner gains a large prospective outlet for its equipment or generation capacity, while Stargate’s first-site construction can be planned around a specific power target.
Second-order effects
- Developers competing for AI-data-center customers will face greater pressure to secure power equipment and generation arrangements alongside land and buildings; powered capacity becomes a key differentiator.
- Crusoe’s delivery risk becomes more visible: its ability to convert the agreement into operating capacity will affect the timetable and economics of the Stargate facility, an issue reflected by the scale of the Abilene site’s multibuilding construction effort.
Third-order effects
- If AI developers continue to procure generation directly, data-center providers will increasingly operate like infrastructure integrators, combining energy procurement, construction, and compute delivery rather than simply leasing server space.
- The shift could favor well-capitalized providers able to lock in long-duration power resources, while making turbine availability, interconnection, and project execution durable constraints on AI expansion.
The trend: AI infrastructure is being built around secured power capacity, pushing compute providers deeper into utility-scale energy development.