Microsoft Working On AI Processor: Expected Launch At Ignite 2023 Conference

In November, Microsoft invites you to the developer and IT expert conference Ignite 2023. A surprise is expected this year – because, according to insiders, the company wants to present its first AI processor.

Generative AI processor planned by Microsoft

Microsoft could therefore be close to revealing its plans to produce its own generative AI processors. According to a report from The Information (paywall), the reveal could come as part of Microsoft’s Ignite 2023 developer conference.

The information comes from an unspecified source but supports older rumors about work on an AI chip. However, details about this AI chip have not yet been given.

The Ignite conference, which is aimed at developers, IT professionals, and administrators, among others, will take place in Seattle from November 14th to 17th.

Microsoft is already advertising that a number of important product announcements can be expected at the conference. It’s no secret that Ignite will focus on new things about the Microsoft AI platform. Microsoft is currently working with Nvidia’s H100 GPUs to power its data centers that process AI services such as Bing Chat, Bing Image Creator, or Copilot.

Microsoft drives Nvidia sales

Microsoft and other generative AI companies bought a lot of Nvidia graphics processors for these data servers last year, driving up the company’s sales and stock price in 2023.

The Microsoft management team is now pushing to end the dependence on chips from third parties such as Nvidia.

Codename Athena

Back in April, The Information reported that Microsoft was developing this new internal AI chip, codenamed Athena. At the time it was said that the chip had been in development since 2019 and that up to 300 team members had been working on the project since then.

In May, another report claimed that Microsoft had partnered with Nvidia’s main GPU rival AMD to co-develop the Athena chip. Microsoft is said to have primarily supported AMD financially so that it could produce its own AI chip.