OpenAI, Oracle Accelerate Stargate AI Expansion Across U.S.

OpenAI has partnered with Oracle to expand its Stargate AI project, adding 4.5 gigawatts of capacity to its U.S. AI infrastructure goals, a move that marks a major milestone in the development of next-gen AI computing.

The development pushes the Stargate program past the 5 gigawatt mark, a clear signal of intent in OpenAI’s push to build out a massive, U.S.-based AI backbone.

Stargate is envisioned as a transformative infrastructure initiative, intended to power the compute needs of advanced artificial intelligence models like GPT-4 and beyond. First announced earlier this year, the program aims to create an AI ecosystem capable of operating at national scale.

With the addition of Oracle, the project gains not only immense technological muscle but also access to Oracle Cloud Infrastructure. OpenAI says this partnership will allow it to deploy state-of-the-art AI models faster, with more energy efficiency and lower latency. Some of the training and inference workloads have already begun using Nvidia’s latest server hardware, installed in Texas at an existing Oracle site.

OpenAI estimates that the Stargate initiative will create over 100,000 jobs across construction, energy infrastructure, logistics, and data center operations. This includes both direct employment and indirect roles in supporting industries.

Despite this momentum, not all elements of the Stargate vision have advanced without issues. Reports suggest that a parallel initiative involving SoftBank has been slowed due to internal disagreements over site selection and coordination. However, OpenAI has reaffirmed that the Oracle-driven arm of the project remains firmly on track.

Oracle is expected to invest tens of billions in capital expenditures to build the required infrastructure, betting on massive cloud revenue returns in the years ahead.