The ‘Stargate’ project is not to build a portal that can take people to other planets as in the film and TV series of the same name. It is a supercomputer designed by Microsoft and OpenAI especially to perform tasks related to artificial intelligence.
According to The Information, Microsoft and OpenAI are in the planning phase to build this supercomputer. Site sources added that ‘Stargate’ will use millions of processing chips for AI and that the budget for this machine to become a reality reaches 100 billion dollars.
The ‘Stargate’ will be the largest of the AI computers that Microsoft and OpenAI want to develop over the next six years as a result of the partnership they have established. If all goes well, this supercomputer will go into operation in 2028.
Although the plans are not yet finalized, it is most likely that Microsoft will provide the majority of the necessary 100 billion dollars. Both companies are evaluating potential chip suppliers for ‘Stargate’, which include AMD, Nvidia and also Microsoft itself. The design of this computer may require the development of some innovations such as racks that allow for greater server density and new cooling solutions that are capable of managing the large amount of heat that is produced by the hardware.
The partnership between Microsoft and OpenAI began in 2019. The two companies have already built several AI supercomputers to power the ChatGPT model and other generative artificial intelligence technologies.
Right now, Microsoft has approved an expansion of the Mount Pleasant data center in Wisconsin that will cost a billion dollars. This installation could evolve into a ‘Phase 4’ system, valued at US$10 billion, which could come into operation in 2026.
To put the ‘Stargate’ budget into context, Microsoft’s annual budget for expanding and maintaining the infostructure that runs the Azure service is $50 billion.
