In January 2025, OpenAI launched Stargate, an advanced AI infrastructure platform aimed at securing long-duration, frontier-scale compute capabilities for both training and inference. The initiative plans to invest up to $500 billion over the next four to five years, with $100 billion earmarked for immediate deployment. Stargate is designed to address increasing compute demands driven by larger AI models and enterprise workloads, prioritizing capacity ownership and power-first site selection. Key partnerships with Oracle and SoftBank have facilitated significant progress by the middle of 2025, including developments totaling approximately 8 GW of power capacity and over $450 billion in projected investments across multiple U.S. sites. Additionally, OpenAI’s strategy involves expanding internationally, with plans for a 1 GW compute cluster in the UAE and a potential $25 billion project in Argentina. Ultimately, Stargate is positioning OpenAI as a leader in AI infrastructure, transforming its role from cloud customer to orchestrator of large-scale collaborative projects.
Source link
