AMD introduces Helion rack blueprint for AI agents

AMD introduced Helion, an open rack blueprint for running AI agents, with configurations planned to become available in November 2026. The company says a CPU-only design using 6th Gen EPYC processors can support up to 32,768 CPU cores for agent execution while connecting to separate infrastructure that runs models. An enterprise configuration combines EPYC processors, Instinct MI350P GPUs and general-purpose computing. An integrated token router directs agents' model calls across frontier and open-source models, while partners can tailor processor density, acceleration and cooling to workload requirements. AMD plans to contribute the specification to the Open Compute Project. Rack configurations are still under development with ecosystem partners.