The hardware integrates a built-in token router that directs requests across frontier and open-source AI models to optimize cost effectiveness. This shift toward agentic infrastructure comes as the development of AI agents drives demand for CPU compute capacity beyond traditional GPU requirements.
Wednesday, October 21, 2026
Story
AMD Unveils Helion AI Agent Rack With 32,768 Cores
AI9 days ago
Partners including Dell, HPE, Cisco and Supermicro are supporting a new rack-scale compute design to handle the operational needs of AI agents. AMD will release the Helion server rack in November 2026 as an open design contributed to the Open Compute Project. The system is available in liquid-cooled and air-cooled configurations, offering a CPU-only version with 32,768 cores for hyperscalers and an all-in-one model pairing EPYC CPUs with Instinct MI350P GPUs for businesses.
Looking for more context or analysis?
Tilt chat provides reliable, tailored insights, alerts, and automations.