Crusoe to Build 900MW AI Data Center Campus for Microsoft, Expanding ‘Stargate’ Ambitions

Crusoe has announced plans to add 900 megawatts of new capacity at its Abilene, Texas data center campus, positioning the site to support growing artificial intelligence workloads from Microsoft.

The expansion will sit alongside an existing 1.2-gigawatt facility being developed in partnership with Oracle and OpenAI as part of the $500 billion Stargate initiative unveiled last year.

Originally, Oracle and OpenAI were expected to lease the additional capacity, but those plans did not materialize due to financing and negotiation challenges. Meta had also been linked to the expansion, though Microsoft has now emerged as the primary occupant for the new campus.

The project will include two new data halls and a dedicated on-site power plant capable of delivering 900 megawatts of behind-the-meter energy. Each facility is designed to support up to 336 megawatts of critical IT load, reflecting the increasing power demands of large-scale AI infrastructure.

Chief executive Chase Lochmiller said the addition of on-site power generation is aimed at accelerating the development of AI infrastructure in the United States, enabling faster deployment timelines in a highly competitive market.

Once completed, the Abilene campus is expected to reach a total capacity of 2.1 gigawatts. However, current operational capacity remains limited, with only a portion of the existing infrastructure powered on. Crusoe indicated that the remaining capacity from earlier phases is scheduled to come online through 2026.

Construction on the new campus is still in early stages, with land clearing and site preparation underway. The company expects the facilities supporting Microsoft’s AI workloads to become operational by mid-2027.

While Crusoe has not disclosed details of the power generation technology, industry observers suggest the timeline points to solutions such as natural gas-based systems or fuel cells to meet the high and continuous energy demands.

The expansion highlights the escalating race among technology companies to secure large-scale, energy-intensive infrastructure required to train and deploy advanced AI models, as partnerships and alliances continue to evolve across the sector.