Microsoft introduces AMD alternative to Nvidia AI processors

At its upcoming Build developer conference next week, Microsoft is set to unveil plans for providing its cloud computing clientele with a selection of AMD artificial intelligence chips, positioning them as competitors to Nvidia’s offerings.

Additionally, Microsoft will introduce a preview of its new Cobalt 100 custom processors during the conference.

These clusters of Advanced Micro Devices’ top-tier MI300X AI chips will be available for purchase through Microsoft’s Azure cloud computing service. This move aims to offer customers an alternative to Nvidia’s H100 family of potent graphics processing units (GPUs), which currently dominate the AI market in data centers but can often be challenging to acquire due to high demand.

To build AI models or run applications, companies often need to assemble multiple GPUs into clusters since the data and computational requirements exceed the capabilities of a single processor.

AMD, projecting $4 billion in AI chip revenue for the current year, has asserted that its chips possess the necessary potency to both train and execute large-scale AI models.

In addition to Nvidia’s premier AI chips, Microsoft’s cloud computing division offers access to its proprietary in-house AI chips known as Maia.

Separately, Microsoft revealed that the Cobalt 100 processors, set to be previewed next week, deliver a 40% performance enhancement compared to other processors utilizing Arm Holdings’ technology. Companies like Snowflake have already started integrating them into their systems.

Initially unveiled in November, the Cobalt chips are undergoing testing for utilization in Teams, Microsoft’s enterprise messaging platform. They are strategically positioned to rival Amazon’s in-house Graviton CPUs.

Partner Resources

Popular Right Now

Edgecore Insight Podcast

Ep-1: Navigating the Waters of Sustainability

Others have also read ...