As a part of OpenAI’s $500 billion investment for AI infrastructure, its Project Stargate gets additional data centres in deal with Oracle, in Texas, US.
OpenAI and Oracle have launched the second phase of their AI infrastructure project, announcing plans to add 4.5 gigawatts (GW) of new data centre capacity in the United States. This expansion, part of the Stargate initiative, will generate over 100,000 jobs across construction, operations, and supply chains, boosting America’s standing in the global AI race.
The new capacity adds to Stargate, the first phase in Abilene, Texas, where Oracle has already delivered data centre infrastructure. NVIDIA GB200 racks are installed, and early AI training and inference workloads are underway. When combined with this new agreement, Stargate’s total capacity under development surpasses 5GW, capable of running more than 2 million AI chips.
This milestone strengthens OpenAI’s January pledge at the White House to invest $500 billion in 10GW of AI infrastructure by 2029. With strong backing from Oracle, SoftBank, and other partners, OpenAI expects to exceed its initial target.
Stargate is moving towards full operations while serving as a testbed for OpenAI’s next-generation research. Construction has already brought in workers from over 20 states, including specialised electricians and technicians.
Meanwhile, OpenAI’s partnership with SoftBank is advancing with site assessments aimed at redesigning data centres for improved AI performance and efficiency. Microsoft continues to be a critical cloud provider for OpenAI through the Stargate platform.
Stargate is central to OpenAI’s vision of building a robust AI infrastructure to deliver advanced technology benefits to a wider audience. It also supports the US government’s broader strategy of economic growth and AI-driven reindustrialisation, with the White House recognising AI infrastructure as a key driver of national competitiveness.
With this second phase, Stargate is firmly positioned as a cornerstone of future AI innovation.
