April 6, 2026

Microsoft takes over Texas data center project after OpenAI exit

By:
Dallas Bond

Microsoft has stepped in to take over a major data center construction project in Abilene, Texas, following OpenAI's decision to withdraw from the expansion. The project, which is being developed by Crusoe, marks a notable shift as both Microsoft and OpenAI continue to pursue their artificial intelligence ambitions independently, despite their shared history and proximity.

A Neighboring Presence in Texas

The new facilities, described as "AI factory" buildings, will be built alongside OpenAI and Oracle's existing data center campus in Abilene. The site is one of the largest AI data center complexes in the United States, with Crusoe already having completed two buildings for OpenAI and Oracle, as well as a 350-megawatt gas-fired power plant that supplements the local electricity grid.

Microsoft's addition to the project will include two new buildings and a new on-site power plant capable of generating 900 megawatts. According to Crusoe co-founder and CEO Chase Lochmiller, the facilities will serve as part of "the industrial foundation for American AI - at a velocity the industry has never seen."

This expansion comes after OpenAI recently decided to scale back its plans for further development at the Abilene site. "Our flagship Stargate site is one of the largest AI data center campuses in the United States", said Sachin Katti, OpenAI’s head of compute infrastructure, in a post on X. "We considered expanding it further but ultimately chose to put that additional capacity in other locations."

OpenAI is focusing on other projects across the country, including a data center in Wisconsin being developed in partnership with Oracle. Currently, OpenAI and Oracle are still moving forward with six additional buildings at the Abilene site, which are expected to be completed by the end of this year.

AI Boom Sparks Industry Shifts

The origins of these data centers trace back to an earlier plan to use the site for cryptocurrency mining. However, with the explosion of AI technologies like ChatGPT, developers pivoted to prioritize AI-focused infrastructure. Microsoft, which once served as OpenAI’s exclusive cloud computing provider and still maintains a 27% stake in the company, is now pursuing its own path in AI development.

Despite this divergence, the collaboration has left a lasting footprint on the Abilene campus. The entire site is expected to supply a total of 2.1 gigawatts of computing capacity once all projects are completed.

Environmental Considerations

The rapid expansion of AI infrastructure raises questions about its environmental impact. The on-site power plants rely heavily on gas, a fossil fuel that contributes to greenhouse gas emissions. Speaking last year about the Abilene data center, OpenAI CEO Sam Altman acknowledged the challenges: "We’re burning gas to run this data center", he said. However, he expressed optimism for the future, adding, "in the long trajectory of Stargate, the hope is to rely on many other power sources."

This tension between technological growth and environmental responsibility reflects the broader challenges facing the AI industry as demand for computing power surges.

Conclusion

Microsoft's decision to take over the expansion of the Texas data center underscores the growing competition in the AI sector, even among former collaborators. As both Microsoft and OpenAI chart their own paths, the Abilene site remains a key focal point for the development of next-generation AI technologies, reshaping the landscape of the region and the broader industry.

Read the source

Keywords:
Microsoft,OpenAI,data center,Abilene,AI infrastructure
Free Download

Data Center Construction Labor Trends in 2026

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.

More mission critical construction news

Data Center Construction Timeline: Realistic Schedules from Greenfield to RFS
May 21, 2026

Data Center Construction Timeline: Realistic Schedules from Greenfield to RFS

Timelines from greenfield to RFS: edge 6–12m, enterprise 12–18m, hyperscale 24–36m; plus risks, long-lead items, and workforce tips.
Data Center Construction Services: Scope, Deliverables, and Common Gaps
May 21, 2026

Data Center Construction Services: Scope, Deliverables, and Common Gaps

Covers data center construction phases, key deliverables, common delivery gaps, and practical fixes for timely, reliable builds.
Top Data Center Construction Companies: How to Shortlist the Right Partner
May 21, 2026

Top Data Center Construction Companies: How to Shortlist the Right Partner

A practical framework to shortlist data center construction partners: define specs, vet MEP skills, manage risk, build owner team.
Data Center Construction Management: What Owners Should Expect from Their CM Firm
May 21, 2026

Data Center Construction Management: What Owners Should Expect from Their CM Firm

Owners should expect CM firms to manage planning, vendor coordination, commissioning, procurement, and risk to deliver ready data centers.