Microsoft revealed plans to enhance its cloud computing offerings with AMD’s advanced artificial intelligence (AI) chips, setting the stage for a new level of competition with Nvidia in AI chips space.
This move will be further detailed at Microsoft’s upcoming Build developer conference. Azure, Microsoft’s cloud service, will now feature clusters of AMD’s MI300X AI chips, providing an alternative to Nvidia’s highly sought-after H100 GPUs, which often face supply shortages due to high demand.
AMD projects that its AI chip revenue will reach USD 4 billion this year, showcasing the power and capability of the MI300X chips in training and running extensive AI models. These chips are integral for businesses that require significant computational power, often necessitating the use of multiple GPUs due to the large volumes of data and processing involved.
In addition to integrating AMD chips, Microsoft will preview its new Cobalt 100 custom processors at the conference. These processors, which boast a 40 per cent performance improvement over other Arm-based processors, represent a significant advancement in processing power. Originally announced in November, the Cobalt chips are currently being tested to enhance Microsoft Teams and are positioned to rival Amazon's in-house Graviton CPUs.
Microsoft’s cloud computing unit is also offering access to its own AI chips, named Maia, providing a broad spectrum of high-performance options for its customers. Companies such as Snowflake have already begun utilising the new Cobalt processors, indicating their potential to impact the cloud computing and AI technology sectors significantly.
