OpenAI is planning to take up data center space beyond its primary backer Microsoft, as it looks to aggressively expand its compute capabilities.

The Information reports that CFO Sarah Friar spoke to employees after the company this month raised $6.6 billion in capital, and said that OpenAI would have a greater involvement in finding data centers and AI chips.

Friar previously told shareholders that Microsoft was moving too slowly for the generative AI startup, something that gained further traction after Elon Musk’s xAI rapidly deployed a giant compute cluster in Memphis.

OpenAI has already begun to move beyond Microsoft, announcing a deal in June with Oracle to rent servers from a data center under construction in Abilene, Texas.

While Microsoft was included in the press release, The Information notes that it is not deeply involved in the project - instead, it will technically rent the servers and then provide them to OpenAI, allowing it to still say it is OpenAI's exclusive cloud provider.

The AI company is now in talks to rent the entire Abilene data center, which is expected to come in just below 1GW by mid-2026. Should Oracle be able to unlock more power, the site could grow to 2GW.

OpenAI could also pursue its long-rumored 'Stargate' 5GW data center without Microsoft, if it can find financing for the $100bn+ facility (and find power).

CEO Sam Altman told employees this week that the cost could be driven down with OpenAI's own custom chips, but it's not clear how far along the project is.