OpenAI Partners Nvidia, Microsoft, SoftBank and Others for New ‘Stargate Project’ to Build AI Infrastructure

OpenAI announced the formation of a new company, dubbed The Stargate Project, on Tuesday. The new venture of the firm will focus on building artificial intelligence (AI) infrastructure in the US. The company announced that it would spend a total of $500 billion (roughly Rs. 43 lakh crore) to build the capacity. The development comes after the AI firm suffered several outages in recent times and had to put up a rate limit on its video generation platform Sora to ease the load on servers. Notably, Microsoft also issued a statement highlighting that the venture did not alter the existing deal it had with OpenAI.

OpenAI Announces the Stargate Project

In a blog post, OpenAI announced that The Stargate Project was aimed at both strengthening the AI firm and securing American leadership in AI. The plan is to invest $100 billion (roughly Rs. 8.6 lakh crore) immediately and the rest ($400 billion) over the next four years. The initial equity funders in the venture are SoftBank, OpenAI, Oracle, and MGX. SoftBank and OpenAI will be the lead partners, with the former handling financial responsibility and the latter handling operational responsibility.

On the technology front, OpenAI is partnering with Arm, Microsoft, NVIDIA, and Oracle. Nvidia and OpenAI will build and operate the new AI infrastructure. The post highlighted that the infrastructure building was already taking place in Texas and the company was evaluating other potential sites.

Notably, OpenAI already has a partnership with Microsoft which will continue till 2030. Under the agreement, Microsoft’s Azure cloud service platform has exclusive rights to OpenAI API. Microsoft, in a statement issued on Wednesday, highlighted that all the existing terms of the deal will continue to be in place.

Microsoft also added that OpenAI “recently made a new, large Azure commitment that will continue to support all OpenAI products as well as training.” The tech giant said that it has approved the ChatGPT maker’s ability to build additional capacity, which will primarily be used for research and training of models.

Last year, The Information reported that OpenAI leadership was getting frustrated with Microsoft as the feeling in the company was that Azure was not moving fast enough to scale its servers to meet OpenAI’s computational needs.