Artificial intelligence news.
Straight from the original announcements.

AMD Helion: Open Rack Blueprint Targets Agentic AI Workloads

October 12, 2026, 15:00Source: AMD
AMD Helion: Open Rack Blueprint Targets Agentic AI Workloads
Image: AMD

AMD has introduced AMD Helion, an open rack-scale blueprint for agentic AI infrastructure based on an Open Compute Project spec-design and powered by AMD compute. According to AMD, it lets customers match CPU performance, core density and power with GPU acceleration for the work their AI agents perform.

Key points
  • AMD introduced Helion, an open agentic AI rack blueprint based on an OCP spec and powered by AMD compute.
  • A CPU-only configuration with 6th Gen AMD EPYC supports up to 32,768 CPU cores for hyperscale customers.
  • An all-in-one rack combines AMD EPYC CPU nodes, AMD Instinct MI350P GPU nodes and general-purpose compute.
  • AMD plans to contribute the Helion specification to OCP.
  • Configurations are under development with partners and planned for November 2026.

What AMD Helion is

AMD announced Helion on October 12, 2026, describing it as an agentic AI rack that builds on the open-platform approach of its AMD Helios rackscale solution. The design is meant to give partners and customers flexibility in how they build infrastructure for AI agents.

The rack blueprint is based on an OCP spec and supports open software. It can come in liquid-cooled and air-cooled versions, which, according to AMD, lets customers create virtually any configuration they need. AMD plans to contribute the Helion specification to OCP so the community can develop and adopt it further.

Dan McNamara, senior vice president and general manager of Compute and Enterprise AI at AMD, said: "Agentic AI is not a single workload." He added that it spans personal agents, enterprise applications, code execution and complex workflows, each with different compute requirements.

Two example configurations

AMD gave two examples of what the blueprint can support:

  • CPU-only rack: powered by 6th Gen AMD EPYC processors, supporting up to 32,768 CPU cores. It is designed for hyperscale customers, with dedicated capacity for agent execution and connections to separate model inference infrastructure.
  • All-in-one rack: combines AMD EPYC CPU nodes, AMD Instinct MI350P GPU nodes and general-purpose compute on one platform. AMD positions it as an all-in-one concept for enterprise agentic AI.

According to AMD, Helion also lets customers deploy agentic AI in the IT infrastructure they already have, turning existing investments into a foundation for new AI capabilities. Partners can tailor CPU and GPU configurations to match agent execution, model inference and supporting compute to each deployment's performance, power and cost requirements.

Partner backing and timing

Several hardware makers voiced support in the announcement: Cisco, Dell Technologies, Hon Hai Technology Group, HPE, Lenovo, QCT, Supermicro and Wiwynn. Supermicro said it will bring Helion, which it described as an ORv3 compliant rack-scale architecture, to market through its Data Center Building Block Solutions. Wiwynn said it will contribute rack-level integration and advanced cooling expertise for hyperscaler customers. Dell referred to the AMD Helion Agentic Rack Reference Design and its own Dell PowerRack.

AMD said Helion agentic rack configurations are currently under development with ecosystem partners and are planned to be available in November 2026. The release carries a forward-looking statement caution, noting that timing and expected benefits could differ from current expectations.

What it means

For data center operators and enterprises, Helion points to rack designs where the balance of CPUs and GPUs can be set according to the agent workload rather than fixed in advance. Hyperscalers get a CPU-heavy option that keeps agent execution separate from model inference, while enterprises get a combined CPU and GPU rack. Because the specification is slated for contribution to OCP and tied to multiple system makers, buyers may see several vendor implementations of the same blueprint, though AMD has not yet detailed specific products or pricing.

Related news

Vultr and AMD open 20 MW AI GPU supercluster in Chennai
Chips and infrastructureNovember 9, 2026, 16:24

Vultr and AMD open 20 MW AI GPU supercluster in Chennai

Vultr and AMD have opened a 20 MW supercluster of AMD Instinct GPUs in Chennai, expanding the cloud provider’s stated commitment to India’s AI infrastructure. The…

AMD EPYC CPUs Power Google Cloud’s Confidential TPU 8tx Infrastructure
Chips and infrastructureNovember 9, 2026, 14:00

AMD EPYC CPUs Power Google Cloud’s Confidential TPU 8tx Infrastructure

5th Gen AMD EPYC processors are powering the host infrastructure for Google Cloud TPU 8tx, which AMD describes as the first cloud Tensor Processing Unit system with…

AMD reports record quarter as data center business nearly doubles
Chips and infrastructureNovember 3, 2026, 22:15

AMD reports record quarter as data center business nearly doubles

AMD posted record revenue of $13.4 billion in the third quarter of 2026, while revenue from its data center business increased 97% year over year. The company expects…