According to a recent report by The Information, there is an ongoing collaboration between tech giants Microsoft and OpenAI that is planning a groundbreaking data center project that could build an artificial intelligence supercomputer called “Stargate” by 2028.
The ambitious endeavor, which could cost as much as $100 billion, comes amid a surge in demand for AI data centers capable of handling increasingly intricate tasks driven by the rapid adoption of generative AI technology.
According to the report, talks about the proposal indicate that Microsoft is ready to take control of the financing, which is expected to significantly exceed the expenses associated with the current data center operations.
The anticipated supercomputer, to be based in the United States, is expected to be the flagship of a series of similar projects planned over the next six years, as The Information details.
The project’s roadmap, outlined by Altman and Microsoft, is reported to include five distinct phases, with “Stargate” being the pinnacle of Phase Five. Previously, Microsoft is said to be currently working on a smaller-scale supercomputer, scheduled for release around 2026.
Much of the projected costs for the coming stages are attributed to the acquisition of specialized AI chips, crucial components that provide advanced AI capabilities. According to Nvidia CEO Jensen Huang, these chips range in price from $30,000 to $40,000.
Microsoft’s vision to advance artificial intelligence capabilities was further underlined by the announcement of purpose-built computing chips last year, which complements its ambitious plans for a fresh supercomputing initiative that is expected to include chips from a variety of vendors.
As The Information notes, projected spending for this endeavor could exceed $115 billion, a significant boost over Microsoft’s infrastructure capital spending last year.
(with support from Reuters)
Posted: March 30, 2024 11:05 am EST