Test report DSG-4254 · Rev A · tested October 12, 2026
AI Datacenter InfrastructureDevice under test
AMD Helion: 32,768-Core CPU-Only Rack Targets Agentic AI Workloads
AMD's Helion rack pairs 6th Gen Epyc 'Venice' CPUs with up to 32,768 cores for agentic AI. Configurations arrive November 2026 via Cisco, Dell, HPE and others.
- Read
- 2 min
- Words
- 350
- Node
- 28nm
- Operator
- Marcus Bennett
Spec summary
- AMD announced Helion on October 12 at the OCP Global Summit in San Jose.
- Helion supports up to 32,768 CPU cores using 6th Gen AMD Epyc 'Venice' processors.
- Configurations are expected to be available in November 2026.
- AMD will contribute the Helion specification to the Open Compute Project.
- Partners include Cisco, Dell, Foxconn, HPE, Lenovo, Supermicro, and Wiwynn.
AMD has announced Helion, a CPU-only rack supporting up to 32,768 CPU cores, built for agentic AI workloads and slated to reach ecosystem partners in November 2026. The company unveiled the platform on October 12 at the OCP Global Summit in San Jose.
Helion builds on what AMD calls the "open-platform approach" of its Helios rack-scale solution, launched earlier this year. The rack targets hyperscale customers and runs on 6th Gen AMD Epyc 'Venice' server CPUs.
What can customers configure?
The base configuration is CPU-only, but AMD said users can customize Helion by mixing CPU and AMD Instinct GPU nodes. The goal: match agent execution, model inference, and the performance, power, and cost profile of each deployment.
The rack ships in two thermal variants:
- Liquid-cooled
- Air-cooled
AMD says this flexibility lets users "create virtually any configuration needed" to support their requirements.
Why CPU-only for agentic AI?
AMD positions Helion around the argument that agentic workloads are heterogeneous rather than monolithic. "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, SVP and GM of compute and enterprise AI at AMD.
"To do that takes optimized compute and infrastructure that can adapt to different workloads. With AMD Helion, we are bringing AMD Epyc performance, open software and flexible rack designs together so our partners can build systems around customer needs and make efficient use of available power," McNamara said.
The 32,768-core ceiling gives operators a large general-purpose compute pool for agent orchestration and code execution, with GPU nodes added only where inference throughput demands it.
When will Helion be open and available?
AMD plans to contribute the Helion specification to the Open Compute Project, making it available to the community for further development and adoption. Rack configurations remain under development with ecosystem partners at present.
The partner lineup for the launch includes:
- Cisco
- Dell
- Hon Hai Technology (Foxconn)
- HPE
- Lenovo
- Supermicro
- Wiwynn
AMD expects Helion agentic rack configurations to be available in November 2026.
via Data Center Dynamics (Source)
More from Marcus Bennett
Show full bio
News editor covering regulation, policy and their impact on hardware layer of artificial intelligence at Die Signal.
311 articles
Same lot · LOT-C1C8
- DSG-12373nmAMD Pitches Helion Rack at Agentic AI Data Centers: Up to 32,768 CPU Cores
- DSG-671165nmMicrosoft, AMD Unveil Helios Rack for Azure AI Inference
- DSG-261114nmAMD Launches Ryzen AI Halo Platform and Ryzen AI Max PRO 400 Series
- DSG-124314nmAMD EPYC 9006: Mapping Venice Silicon to Agentic AI Workloads
- DSG-831028nmAMD Buys Taalas to Add Model-Specific Silicon to Instinct Roadmap