OpenAI and Oracle reportedly abandon TX Stargate expansion • The Register

OpenAI and Oracle have reportedly abandoned plans to expand their flagship Stargate datacenter in Abilene, Texas, after negotiations were stalled by financing challenges and OpenAI's inability to forecast demand effectively, according to sources familiar with the matter cited by Bloomberg.

The companies had planned to grow the facility to as much as 2 gigawatts of capacity, up from the 1.2 gigawatt campus currently under development. The expansion's cancellation comes as Meta is reportedly in discussions to lease the untapped and unbuilt capacity from the site's developer Crusoe, with Nvidia having stepped in to broker the deal.

According to Bloomberg, Nvidia put down a $150 million deposit on the future capacity before approaching Meta about potentially moving in. When reached for comment, Crusoe declined to provide details on the negotiations.

This development marks a setback for the $500 billion Stargate initiative announced just over a year ago, which aims to deploy massive AI compute infrastructure across the United States. Despite this cancellation, OpenAI's broader partnership with Oracle to deploy 4.5 gigawatts of compute capacity remains on track, with the contract valued at $300 billion over its lifetime.

The financing challenges highlight the massive capital requirements for AI infrastructure expansion. This week, Oracle announced plans to raise an additional $50 billion in debt and equity to finance its datacenter aspirations, though this funding won't arrive in time to save the Abilene expansion.

The datacenter capacity race continues to intensify among major tech companies. Meta CEO Mark Zuckerberg has shown particular enthusiasm for securing compute resources, with the company announcing during its Q4 earnings call in January its intention to invest up to $135 billion in capital expenditures focused on GPU compute capacity.

Combined, the eight largest hyperscalers—Google, Amazon, Meta, Microsoft, Oracle, Tencent, Alibaba, and Baidu—are expected to spend a collective $710 billion in 2026 alone, much of it on datacenters and the GPUs and other infrastructure that go into them.

For OpenAI, the cancellation creates a challenge in meeting its compute requirements. The company needs to find capacity for 5 gigawatts worth of GPUs to claim the $30 billion incentive offered by Nvidia as part of a $110 billion funding round announced last week in collaboration with SoftBank and Amazon.

The Stargate expansion's cancellation underscores the volatile nature of AI infrastructure planning, where demand forecasting remains uncertain and financing massive projects requires careful timing and coordination among multiple stakeholders.