AMD Launches Helion AI Compute Rack
AMD announced its Helion rack solution for agent-based artificial intelligence workloads. The system is built on OCP specifications and offers flexible combinations of CPU and GPU resources.
Advanced Micro Devices (AMD) has unveiled the Helion rack solution, designed for agent-based artificial intelligence computing tasks. The system leverages Open Compute Project (OCP) design specifications and AMD's Helios rack platform principles.
The Helion architecture allows partners and customers to configure a mix of CPU and GPU resources tailored to their specific AI workload requirements. A pure CPU configuration, utilizing AMD's sixth-generation EPYC processors, can scale up to 32,768 cores.
AMD intends to contribute the Helion rack specification to the OCP community. This move aims to foster further development and adoption of the platform within the broader industry. The company is collaborating with ecosystem partners to develop the Helion rack.
Commercial availability of the AMD Helion rack solution is planned for November 2026. The initiative underscores AMD's focus on providing scalable compute solutions for the growing AI sector.