OpenAI, Oracle, SoftBank Unveil 5 New Stargate AI Data Center Sites

OpenAI, Oracle, and SoftBank announced five new U.S. sites under the Stargate AI data center program, accelerating their push to deploy 10 gigawatts of AI computing capacity. The combined capacity from these new builds—together with the flagship campus in Abilene, Texas, and existing projects with CoreWeave—brings Stargate to nearly 7 GW and over $400 billion in committed investment, with a full $500 billion expected by the end of 2025, ahead of schedule.

Three of the new sites fall under OpenAI’s partnership with Oracle, which in July agreed to develop up to 4.5 GW of Stargate capacity worth more than $300 billion over five years. These include new facilities in Shackelford County, Texas, and Doña Ana County, New Mexico, plus a yet-to-be-disclosed Midwest location, along with a 600 MW expansion near Abilene. The Oracle-backed facilities can collectively deliver more than 5.5 GW and are expected to generate over 25,000 onsite jobs.

The other two sites are joint projects between OpenAI and SoftBank. A Lordstown, Ohio, site is already under construction and due online in 2026, while a Milam County, Texas, project with SB Energy will provide powered land for rapid buildout. Both sites are designed to scale to 1.5 GW within 18 months, with potential for further growth. Together, the five sites mark the first phase of selections from more than 300 proposals reviewed across 30 states.

• Nearly 7 GW of Stargate capacity committed, with $400B investment to date

• $500B, 10 GW U.S. target expected by end of 2025, ahead of schedule

• Oracle partnership adds over 5.5 GW across three new sites and one expansion

• SoftBank partnership adds 1.5 GW capacity in Ohio and Texas

• More than 25,000 onsite jobs expected, with tens of thousands more indirectly

“AI can only fulfill its promise if we build the compute to power it. That compute is the key to ensuring everyone can benefit from AI and to unlocking future breakthroughs,” said Sam Altman, CEO of OpenAI.

🌐 Analysis: OpenAI is advancing at unprecedented scale in AI infrastructure, positioning itself at the center of global compute expansion. In June 2025, the company confirmed it had begun training workloads on NVIDIA’s new GB200 systems at its flagship Stargate campus in Abilene, Texas — one of the first large-scale deployments of NVIDIA’s Blackwell architecture. This milestone demonstrated not only early access to cutting-edge silicon but also the speed with which OpenAI is operationalizing next-generation compute for AI research.

The Stargate program, announced in January 2025 with Oracle and SoftBank, is on a clear path toward nearly 7 GW of planned capacity, with $400 billion in investments already committed. The partnership with Oracle represents more than $300 billion over five years, enabling multi-gigawatt facilities in Texas, New Mexico, and the Midwest. SoftBank, through SB Energy, adds a separate 1.5 GW tranche of rapid-build sites in Ohio and Texas. In parallel, OpenAI continues to collaborate with CoreWeave on distributed compute capacity to complement the Stargate mega-campuses.

Internationally, OpenAI has been extending its footprint with strategic projects in both the Middle East and Europe. In Abu Dhabi, it is developing a hyperscale AI campus aligned with sovereign energy-backed infrastructure initiatives. In the UK, OpenAI has partnered with Nscale to support the Stargate-linked “Stargate UK” supercomputer program, designed to deliver national-scale compute capacity for AI research and industry. These international efforts, together with U.S. expansion and NVIDIA’s global supply chain, highlight OpenAI’s strategy to establish a geographically diverse and resilient compute backbone capable of supporting next-generation model training and deployment.