Microsoft and OpenAI invest $100 billion in AI supercomputer and new data center – Computers

Generative AI technology continues to evolve at a rapid pace and the increase in demand for models requires the construction of dedicated data centers. A Microsoft and OpenAI are working on a new data center project, with an investment estimated at more than $100 billion. This includes a supercomputer dedicated to artificial intelligence called Stargate, scheduled for launch in 2028.

According to The Information, the supercomputer could have millions of specialized chips to power OpenAI’s artificial intelligence, information obtained unofficially according to three people involved in planning the proposal. A Microsoft is expected to be responsible for financing the project, at an estimate of 100 times the cost of today’s largest data centers. This is a demonstration of the computing needs for AI in the coming years.

Stargate promises to be the largest supercomputer among proposals from different companies, scheduled for the next six years. Microsoft and OpenAI’s plans are to spread supercomputers across five phases, with Stargate being the fifth phase. Before that, Microsoft is working on a smaller, fourth-phase supercomputer for OpenAI, which will be released in 2026. And they are currently halfway through the third phase.

Second is advanced, the next two phases require the acquisition of the necessary AI chips. Chips that have been sold at high prices. Nvidia recently unveiled the Blackwell B200 AI chip, which is priced between 30 and 40 thousand dollars. So that companies are not tied to just one manufacturer, the new project has a design capable of supporting chips from different suppliers, according to Reuters.

O budget could exceed 115 billion dollars, representing more than triple what Microsoft spent last year on investment in serversbuildings and other equipment, reports the news agency.