Microsoft (NASDAQ:MSFT) revealed its intention on Thursday to provide its cloud computing clients with a selection of AMD artificial intelligence chips as an alternative to Nvidia (NASDAQ:NVDA) processors. Further details regarding this initiative are scheduled to be unveiled at the upcoming Build developer conference next week.
The tech giant also announced plans to unveil a preview of its new Cobalt 100 custom processors during the conference. These processors, known as Microsoft’s clusters of Advanced Micro Devices’ flagship MI300X AI chips, will be available through its Azure cloud computing service. They will provide customers with an alternative to Nvidia’s H100 family of powerful GPUs, which currently dominate the data center chip market for AI but can be challenging to acquire due to high demand.
Companies typically need to cluster multiple GPUs together to build AI models or run applications, as the data and computation may not fit on a single processor. AMD anticipates generating $4 billion in AI chip revenue this year, touting the power of its chips to train and run large AI models effectively.
In addition to offering Nvidia’s top-shelf AI chips, Microsoft’s cloud computing unit provides access to its proprietary in-house AI chips named Maia.
Separately, Microsoft announced that its Cobalt 100 processors, set to be previewed next week, offer a 40% improvement in performance over other processors based on Arm Holdings’ technology. Several companies, including Snowflake, have already begun using these processors.
The Cobalt chips, announced in November, are currently undergoing testing to power Teams, Microsoft’s messaging tool for businesses. They are positioned to compete with Amazon.com’s in-house Graviton CPUs.
Featured Image: Freepik
