Photo by Jeswin Thomas on Pexels

Oracle is making a monumental $40 billion investment in Nvidia chips to power a cutting-edge data center dedicated to artificial intelligence, specifically supporting OpenAI’s ambitious computing needs. The facility, located in Abilene, Texas, marks a significant step in addressing the surging demand for AI processing capabilities.

The Texas-based “Stargate” data center, a key component of a larger AI infrastructure plan involving OpenAI and SoftBank, is projected to deliver 1.2 gigawatts of computing power, potentially making it one of the world’s largest. Oracle plans to acquire approximately 400,000 of Nvidia’s advanced GB200 chips, designed for AI model training and execution, and lease the computing power to OpenAI.

Crusoe and Blue Owl Capital own the site and have secured $15 billion in funding for the project, with Oracle committing to a 15-year lease. JPMorgan Chase is providing the bulk of the debt financing. This initiative mirrors similar expansions in the AI computing landscape, including Elon Musk’s plan to deploy approximately 1 million Nvidia chips in Memphis, Tennessee, and Amazon’s development of a separate data center in Virginia exceeding 1 gigawatt.

The “Stargate” project represents a strategic move for OpenAI to diversify its infrastructure and reduce reliance on Microsoft. Following the termination of their exclusive partnership, OpenAI and Microsoft are currently negotiating licensing rights for OpenAI’s models. The ultimate goal of the “Stargate” initiative is to secure substantial funding, potentially reaching $500 billion over four years, to support further data center development across the United States. Oracle and MGX have also pledged funds to the initiative, which includes an additional “Stargate” facility in Abu Dhabi spanning 10 square miles with 5 gigawatts of data center capacity.