According to 1M AI Odaily, OpenAI has made significant adjustments to its Stargate computing infrastructure project, abandoning its self-built data centers and instead adopting third-party cloud services such as AWS and Google Cloud to provide computing power. The company has lowered its total computing power expenditure forecast for 2030 from $1.4 trillion to approximately $600 billion.
This adjustment is primarily due to financing pressures. Since its inception in 2025, the Stargate project has progressed slowly and has failed to secure large-scale construction financing. OpenAI has withdrawn from related expansion negotiations in Texas and reached a new cooperation agreement with SoftBank, whereby SoftBank will be responsible for energy and infrastructure development, while OpenAI will acquire computing resources through long-term leases.
In addition, OpenAI is shifting its computing power deployment to the NVIDIA Vera Rubin platform, aiming to achieve its first gigawatt-level computing power scale in the second half of 2026.
