OpenAI, Oracle, and SoftBank add five Stargate AI data centers, pushing capacity toward 7 GW
OpenAI, Oracle, and SoftBank unveiled five new U.S. data center sites under the Stargate program, moving the $500 billion initiative closer to its 10-gigawatt compute goal and lifting planned capacity to nearly 7 GW over the next three years. The new locations span Texas, New Mexico, Ohio, and an undisclosed Midwest site.
The announcement
- Five sites: Shackelford County (TX), Doña Ana County (NM), Milam County (TX), Lordstown (OH), plus a Midwest site to be named.
- Investment to date surpasses $400B; total commitment remains $500B targeting 10 GW.
- Expected to create more than 25,000 onsite jobs.
- Abilene, TX flagship campus is already online and expanding.
How the buildout works
- Oracle will co-develop three sites and a potential 600 MW expansion near Abilene; OpenAI and Oracle’s July agreement covers up to 4.5 GW of additional capacity.
- SoftBank and OpenAI will co-develop two sites that can scale to 1.5 GW within 18 months, including a Lordstown facility slated to go operational next year.
- Early workloads at Abilene are running on Oracle Cloud Infrastructure with NVIDIA GB200 racks already delivered.
- NVIDIA is positioned as a key supplier, with reports indicating up to $100B in chip commitments for Stargate.
Why it matters
- More compute for generative AI products (e.g., ChatGPT) and enterprise AI agents, easing capacity constraints and latency.
- Signals intensifying competition across cloud and chip supply chains and could influence regional energy planning.
- The buildout also broadens OpenAI’s infrastructure footprint beyond Microsoft’s Azure while keeping multiple partners in the mix.
What’s next
- Additional U.S. sites are expected as the consortium targets securing the full 10 GW, $500B commitment by end of 2025—described as ahead of schedule.
- Construction timelines vary by site; Oracle- and SoftBank-led campuses begin scaling through 2026, pending local approvals and power availability.
In short, Stargate’s latest wave marks one of the most aggressive AI infrastructure expansions to date—pairing cloud capacity, next-gen GPUs, and multi-partner financing to meet surging demand for AI training and inference across industries.