OpenAI and Oracle announce 4.5 GW expansion of Stargate AI data centers in the US

OpenAI and Oracle expand Stargate with a record-breaking 4.5 GW of new AI data center capacity. The partnership ushers in a new era of supercomputing for artificial intelligence.

OpenAI and Oracle have signed a historic agreement to build an additional 4.5 gigawatts of AI-focused data center capacity across the United States. This massive infrastructure project, now known as Stargate, places the US at the forefront of global AI innovation and lays the foundation for future generations of large language models.

The Stargate project enters a new phase with unprecedented scale and ambition

With the announcement of 4.5 GW of new data center capacity, the Stargate partnership now surpasses 5 GW under development—enough to run over two million specialized accelerator chips. This expansion rivals the power demand of entire cities and instantly ranks Stargate among the world’s largest cloud infrastructure projects. The first site, Stargate I in Abilene, Texas, began construction in early 2025 and is already running early AI workloads, with more sites set to come online through 2026.

The collaboration is designed to accelerate the future of AI research and model training

Oracle Cloud Infrastructure (OCI) will lead the design and operation of the new campuses, providing OpenAI with priority access to racks equipped with Nvidia’s latest GB200 Grace Blackwell superchips. These chips, purpose-built for next-generation AI, will enable OpenAI to train, deploy, and experiment with increasingly sophisticated models—including the future GPT-5 and its successors—at a scale previously unthinkable.

Economic impact and job creation are central to the Stargate vision

OpenAI estimates that this new phase will generate more than 100,000 jobs across the US, including construction, operations, manufacturing, and services. By investing directly in national infrastructure, OpenAI and Oracle are supporting local economies and helping to position the US as a world leader in AI hardware and cloud services.

A multi-vendor ecosystem brings together major partners for global reach

Stargate is not just an OpenAI-Oracle initiative. Microsoft, OpenAI’s primary cloud provider, remains deeply involved—interconnecting Azure regions, sharing safety tools, and supporting the development of a multi-tenant AI infrastructure. Additional partners like SoftBank and CoreWeave contribute GPU capacity and are scouting sites for future expansions, reinforcing the vision of a broad, collaborative AI platform that spans multiple technology leaders.

The Stargate expansion signals a strategic shift to dedicated AI supercomputing campuses

Industry analysts note that OpenAI’s aggressive infrastructure strategy signals a new era, where the world’s most advanced models are trained not on rented, shared capacity, but on dedicated, purpose-built campuses. For context, the total capacity announced for Stargate is already comparable to the 14 GW Google has deployed across all its global data centers. This positions OpenAI to meet surging global demand for generative AI and to push the boundaries of what is computationally possible.

Looking ahead: the roadmap to 10 GW and international ambitions

In January 2025, OpenAI pledged to invest $500 billion in building 10 GW of AI infrastructure by 2029. The new agreement with Oracle makes that ambitious target attainable well ahead of schedule. OpenAI also hints at further site announcements and the possible launch of “OpenAI for Countries”—joint ventures to bring Stargate-like capabilities to governments worldwide.

With Stargate, OpenAI and Oracle are building the backbone for the next decade of artificial intelligence—combining scale, innovation, and economic opportunity to redefine what’s possible in the age of AI.

________

FOLLOW US FOR MORE.

DATA STUDIOS

datastudios.org