Oracle and OpenAI end plans to expand flagship data centre
Published Sat, Mar 7, 2026 · 10:32 AM
[SAN FRANCISCO] Oracle and OpenAI have scrapped plans to expand a flagship artificial intelligence (AI) data centre in Texas after negotiations dragged on over financing and OpenAI’s changing needs.
The collapsed talks created an opening for Meta Platforms to step in and consider leasing the planned expansion site in Abilene, Texas, from developer Crusoe, according to sources familiar with the matter. Nvidia, the leading AI chipmaker, helped facilitate Meta’s discussions with the developer, said the sources, who asked not to be identified because the talks are private.
The shifting plans underscore the complexity of building out AI data centres, which are expected to cost in the tens of billions of US dollars and require cooperation from a wide swath of partners.
The campus being developed by Crusoe in Abilene is part of the highly publicised Stargate project, which was announced last year at the White House with US President Donald Trump. While the 1,000-acre site continues to be built, and several parts are up and running, Oracle and OpenAI elected not to go forward with tentative plans to lease a large expansion, the sources said.
Oracle and OpenAI are using Nvidia’s AI semiconductors at the Stargate site. With Crusoe seeking a tenant, Nvidia became involved to ensure its products would still fill the expanded data centre rather than that of rival Advanced Micro Devices (AMD), said the sources. Nvidia paid a US$150 million deposit to Crusoe and began helping court Meta as a tenant for the expansion, the sources said.
Oracle agreed last July to develop 4.5 gigawatts of data centre capacity for OpenAI. That deal remains on track, and the companies have announced a number of projects in other locations, such as one near Detroit owned by Related Digital.
Navigate Asia in
a new global order
Get the insights delivered to your inbox.
Shares of Oracle gave up earlier gains and slipped 1.2 per cent to US$152.96 at the close on Friday (Mar 6) in New York. Other companies tied to the AI infrastructure build-out, such as CoreWeave, AMD and Nvidia, also declined on the news.
Intense computing power needed to train and deploy AI models has led to a boom in data centre projects of unprecedented scale. Oracle has transformed its business to focus on filling these demands for clients such as OpenAI and Elon Musk’s xAI. Social networking giant Meta has been spending lavishly on AI infrastructure for its products, and has projected capital expenditures of as much as US$135 billion in 2026 alone. Nvidia is the leading maker of chips for AI workloads.
The Crusoe-owned data centre in Abilene is among the highest-profile yet announced. Oracle has been rapidly filling buildings on the site with servers, which are used by OpenAI to train and deploy its products. Oracle, Crusoe, and OpenAI had been talking since the middle of 2025 about expanding the facility from 1.2 gigawatts to about two gigawatts. A gigawatt is akin to the capacity from one nuclear reactor and can provide electricity at any one point to roughly 750,000 houses.
The negotiations stretched on and were complicated by financing and OpenAI’s often-changing demand forecasting before they fell apart, the sources said.
In addition, relations between Oracle and Crusoe have been strained by reliability issues at the site. Earlier this year, data centre buildings went offline for days due to winter weather affecting some of the liquid cooling machinery, according to sources familiar with the incident.
Both companies said their relationship remains strong and the Oracle-leased site in Abilene is progressing quickly.
“We are very proud of our relationship and our progress in bringing capacity online,” Oracle said.
“Together, Crusoe and Oracle are operating in lockstep to deliver one of the world’s largest AI factories in Abilene,” Crusoe said. “Our collaboration can deliver massive-scale infrastructure faster than anyone else in the industry.”
Meta and OpenAI declined to comment. Nvidia did not respond to requests for comment.
In a social media post after the story was published, OpenAI infrastructure executive Sachin Katti wrote, “our flagship Stargate site is one of the largest AI data centre campuses in the United States. We considered expanding it further, but ultimately chose to put that additional capacity in other locations”.
Negotiations between Meta and Crusoe for the expansion of the Abilene location remain active and could change, the sources said. Meta is currently working on several large data centres in Louisiana and Indiana. Last month, Meta struck a deal to deploy six gigawatts worth of gear from AMD. BLOOMBERG
Decoding Asia newsletter: your guide to navigating Asia in a new global order. Sign up here to get Decoding Asia newsletter. Delivered to your inbox. Free.