AMD Built a Rack Server Around What AI Agents Actually Need to Run

AMD introduced the Helion agentic AI rack on October 12, 2026, positioning the product as purpose-built infrastructure for the distinct compute demands of agentic workloads. The announcement was made at the OCP Global Summit in San Jose.

According to the company, the design premise behind Helion is that agentic AI does not map cleanly onto existing infrastructure categories. "Agentic AI is not a single workload. It spans personal agents, enterprise applications, code execution and complex workflows, each with different compute requirements," said Dan McNamara, Senior Vice President and General Manager of Compute and Enterprise AI at AMD.

The rack is built to let customers and partners configure compute accordingly rather than applying a fixed hardware formula across different deployment types.

Helion is an OCP spec-based rack blueprint available in liquid- and air-cooled versions. Two configurations define its range, according to the press release.

The first is a CPU-only rack powered by 6th Gen AMD EPYC processors, supporting up to 32,768 CPU cores, designed for hyperscale customers who want dedicated capacity for agent execution while connecting to separate model inference infrastructure.

The second is an all-in-one rack combining AMD EPYC CPU nodes, AMD Instinct MI350P GPU nodes, and general-purpose compute on a single platform, designed for enterprise agentic AI deployments.

The platform includes an integrated token router that directs agent model calls across frontier and open-source models. AMD said it plans to contribute the Helion specification to OCP for broader community development.

Eight partners have confirmed support for the platform at launch: Cisco, Dell Technologies, Hon Hai Technology Group, HPE, Lenovo, QCT, Supermicro, and Wiwynn. Rack configurations are currently under development and are planned to be available in November 2026.

Dell described Helion as extending its PowerRack platform to agentic AI, with configurations that combine AMD's rack-scale architecture with Dell's engineering alongside networking and storage options.

HPE framed its involvement as a continuation of its longstanding partnership with AMD on open standards, noting support for both air-cooled and direct liquid-cooled networking. Supermicro said it would bring Helion to market through its Data Center Building Block Solutions.

Helion follows AMD's Helios rackscale solution and represents the company's bid to bring the same open-platform approach to the agentic infrastructure market, where Nvidia currently leads with its own rack reference designs including the NVL72.