Microsoft Deploys AMD Helios AI Systems in Azure Centers
Microsoft will integrate AMD Helios rack-scale AI systems into Azure data centers later this year to support frontier AI model inference and services.
Microsoft announced on Monday that it will deploy AMD's Helios rack-scale AI system across its Azure data centers later this year. The integration aims to provide the performance and scale required for frontier AI model inference and Azure AI services. Alongside the Helios deployment, Microsoft is launching two new Azure computing instances powered by AMD Venice CPUs, specifically designed for agentic AI and semiconductor design.
Helios represents AMD's first integrated AI platform, combining CPUs, GPUs, software, and networking to challenge the market dominance of Nvidia's Grace Blackwell and Vera Rubin systems. Other early adopters of the platform include OpenAI, Meta, Oracle, and Tata Consultancy Services. AMD intends to erode Nvidia's 95% share of the data center GPU market by focusing on a lower cost per token for customers.
Financial projections from AMD indicate that the Helios processor will drive the majority of tens of billions of dollars in data center AI revenue starting in 2027.