OpenAI and Oracle have announced a significant expansion of the U.S. data center capacity for the Stargate project, increasing it by 4.5 gigawatts to a total of just over 5 gigawatts. This move is a crucial step towards OpenAI’s ambitious goal of achieving 10 gigawatts of data center capacity by 2029.
The Stargate initiative is designed to establish the U.S. as a dominant force in global AI development, aligning with a broader national push. The project has garnered significant attention, with President Trump publicly endorsing a $500 billion target for the project by 2029, as reported by The Wall Street Journal.
Discussions regarding Stargate began as early as 2022, but the project was formally announced in January 2025. It involves multiple stakeholders from the technology and investment sectors, with OpenAI providing leadership, AI model development, and technological support. Oracle’s latest agreement with OpenAI is crucial for realizing the project’s goals. SoftBank manages investments and funding, with both OpenAI and SoftBank pledging $19 billion to support Stargate’s development. Crusoe, a Denver-based startup, is responsible for the construction and buildout of the data centers. MGX, an AI investment firm from the UAE, has contributed approximately $7 billion to the project.
Beyond OpenAI and Oracle, other companies are contributing critical technology and support to the Stargate project. ARM (ARM Holdings) designs energy-efficient CPU architectures, while Microsoft serves as a key technology and cloud computing provider. NVIDIA plans to deploy no less than 64,000 GPUs for the project by 2026.
According to OpenAI’s blog post, the additional 4.5 GW of data center capacity announced will create over 100,000 jobs across construction and operations roles in the US. With billions in funding, a coalition of global tech leaders, and government alignment, the Stargate Project is positioned as one of the most ambitious AI infrastructure initiatives in U.S. history. Its scale is intended to not only accelerate AI adoption domestically but also to influence the global AI landscape.
