OpenAI and Oracle expand Stargate AI infrastructure with 4.5 GW data centre additional capacity

OpenAI and Oracle are jointly developing an additional 4.5 gigawatts (GW) of data centre capacity in the U.S., significantly expanding their ongoing Stargate initiative. This move brings total Stargate capacity under development to over 5 GW, supporting OpenAI’s long-term goal of building 10 GW of AI infrastructure in the U.S. within four years.

“We have signed a deal for an additional 4.5 gigawatts of capacity with Oracle as part of Stargate. Easy to throw around numbers, but this is a _gigantic_ infrastructure project,” OpenAI CEO Sam Altman said in a post on X.

The companies did not disclose locations or funding specifics for the new sites. Reportedly OpenAI is expected to incur a cost of around $30 billion annually, according to WSJ. However, the initiative builds on OpenAI’s existing Stargate I facility in Abilene, Texas, where early operations are already underway.

In its Tuesday evening announcement, OpenAI stated that the expanded 4.5 GW project will generate over 1,00,000 U.S. jobs across construction, operations, and ancillary sectors. This includes full-time data centre roles and short-term jobs such as electricians, equipment operators, and technicians. The Stargate I site alone has already employed workers from more than 20 states of the U.S.

In total, the 5+ GW capacity under development will be powered by more than 2 million AI chips. This infrastructure is intended to support training and inference workloads for OpenAI’s next-generation AI models.

What is Stargate and what has been the progress so far?

Stargate I, the first site under the initiative, is already partially operational. Oracle began delivering Nvidia GB200 racks last month, and OpenAI has started using the facility for internal AI workloads.

The partnership with Oracle complements OpenAI’s collaboration with Japanese tech investor SoftBank. Site assessments are in progress under that partnership, with an emphasis on new approaches to AI data centre design.

“Stargate is OpenAI’s overarching AI infrastructure platform—encompassing ongoing data centre partnerships with Oracle, SoftBank and CoreWeave and international investments in U.S. infrastructure through OpenAI for Countries⁠. Microsoft will continue to provide cloud services for OpenAI, including through Stargate,” OpenAI said in a statement.

The expansion reflects growing demand for AI infrastructure amid heightened geopolitical focus on AI leadership. Microsoft-backed OpenAI is one of several U.S. tech firms rapidly building out compute capacity to support generative AI tools such as ChatGPT and Copilot.

Stargate is OpenAI’s central AI infrastructure program. Some of the partners for this project include Oracle, SoftBank, and CoreWeave, aside from international investment into the country’s infrastructure. OpenAI now, expects to exceed its original 10 GW, $500 billion commitment made in February.

“We now expect to exceed our initial commitment thanks to strong momentum with partners including Oracle and SoftBank,” the company added.