Microsoft to Deploy AMD Helios AI Platform on Azure for AI Inference
Microsoft has announced the integration of AMD's next-generation Helios AI platform and EPYC Venice processors into its Azure cloud infrastructure. This deployment will power three new virtual machine series optimized for AI inference, data processing, and electronic design automation (EDA). This partnership strengthens Azure's high-performance computing capabilities and provides cloud customers with powerful alternatives to Nvidia's dominant hardware for demanding AI workloads. It highlights the intensifying competition in the AI hardware market as cloud providers diversify their chip architectures. The liquid-cooled AMD Helios rack-scale platform integrates up to 72 Instinct MI455X accelerators and 256-core EPYC Venice processors, delivering up to 31TB of HBM4 memory and 2900 PFLOPS of FP4 compute power. The new Azure VMs include the HDv2 for data processing, HXv2 for EDA, and ND MI455X v7 for AI inference.
## BACKGROUND
Electronic Design Automation (EDA) refers to software tools used for designing complex electronic systems, such as integrated circuits and printed circuit boards, which require massive computing power. AMD's Instinct MI455X is a next-generation GPU accelerator built on the CDNA architecture, designed specifically to handle frontier-scale AI training and inference workloads.