AMD unveils “Helios” rack-scale AI platform at OCP Summit, built on Meta’s open rack standard

has officially introduced its new rack-scale , “Helios,” at the Open Compute Project (OCP) Global Summit. Designed to meet the demands of next-generation AI workloads, Helios is built on the Open Rack Wide (ORW) specification contributed to OCP by , and integrates AMD’s full stack of high-performance computing technologies.

A new benchmark for Open AI infrastructure

Helios combines AMD Instinct GPUs, EPYC CPUs, and Pensando networking to deliver scalable, energy-efficient performance for gigawatt-scale data centers. The platform is optimized for power, cooling, and serviceability, and supports open compute standards including OCP DC-MHS, UALink, and Ultra Ethernet Consortium (UEC) architectures.

“Open collaboration is key to scaling AI efficiently,” said Forrest Norrod, EVP and GM of AMD’s Data Center Solutions Group. “With Helios, we’re turning open standards into real, deployable systems.”

Helios features a double-wide rack layout, quick-disconnect liquid cooling, and multi-path Ethernet resiliency, making it ideal for mission-critical AI and HPC workloads. As a reference design, it enables OEMs, ODMs, and hyperscalers to rapidly deploy and customize open AI systems, reducing time-to-market and improving interoperability.