AMD unveils Helion AI agent rack solution, with each rack supporting up to 32,768 CPU cores.

Beating AI News (from Dongcha) – AMD today announced the launch of the AMD Helion Agent AI Rack Solution. Built on an open infrastructure design compliant with Open Compute Project (OCP) specifications, the solution is engineered to meet the computing power demands of large-scale AI agent tasks, supporting flexible combinations of CPUs and GPUs. It offers two primary configurations: The first is a CPU-only rack powered by 6th Gen AMD EPYC processors, supporting up to 32,768 CPU cores, designed for hyperscale deployments, providing dedicated computing power for agent tasks and connecting to separate model inference infrastructure. The second is an integrated rack combining AMD EPYC CPUs, AMD Instinct MI350P GPUs, and general-purpose compute nodes, targeted at enterprise-grade agent AI applications. AMD states that Helion supports both liquid and air cooling configurations, allowing customers to adjust CPU and GPU resources based on their workload, performance, power consumption, and cost needs. The solution also integrates open-source software and hardware plus a built-in token router, enabling the distribution of agent model calls to both frontier and open-source models. AMD plans to submit the Helion specifications to the OCP for further community development and adoption. Currently, the solution is being developed in collaboration with ecosystem partners, with a launch scheduled for November 2026.