Microsoft joins Meta and OpenAI to deploy AMD Helios AI racks

Microsoft is expanding Azure’s AI infrastructure with AMD’s Helios rack-scale systems and next-generation Instinct and EPYC chips for large-scale training, inference and production workloads.

Summary

Microsoft is expanding its partnership with AMD to deploy the chipmaker’s Helios rack-scale AI systems in Azure and to use next-generation AMD Instinct accelerators and EPYC server processors more broadly in its infrastructure. AMD has previously identified Microsoft, Meta, OpenAI and Oracle as Helios customers, with shipments expected later this year, though the companies have not disclosed Microsoft-specific deployment timing, volumes, financial terms or operational targets.

Terms & Concepts
  • Helios: AMD’s rack-scale AI infrastructure design combining Instinct accelerators, EPYC processors, Pensando networking and ROCm software for large-scale AI workloads.
  • Instinct: AMD’s accelerator chip family for AI training and inference workloads.
  • EPYC: AMD’s server processor family for data center and cloud infrastructure.