AMD and Hewlett Packard Enterprise expand AI infrastructure collaboration
AMD and Hewlett Packard Enterprise (HPE) revealed a broadening of their cooperative efforts focusing on Artificial Intelligence (AI) infrastructure, specifically adopting AMD’s “Helios” rack-scale architecture. HPE will integrate this AI platform with purpose-built networking technology and plans to launch the architecture globally in 2026.
This collaboration targets operational and strategic optimization for organizations managing extensive AI workloads. The “Helios” platform is designed to enhance deployment efficiency, scalability, and flexibility in AI compute infrastructure, which is relevant to research institutions and cloud service businesses.
The “Helios” system combines AMD EPYC CPUs, Instinct GPUs, and Pensando networking products. It supports up to 2.9 exaFLOPS of FP4 performance per rack and utilizes open standards, including the ROCm software stack, to coordinate high-bandwidth, low-latency connectivity across AI clusters. HPE contributed a scale-up Ethernet switch developed alongside Broadcom, which leverages the Ultra Accelerator Link over Ethernet (UALoE) standard for optimized data flow.
The partners have implemented the “Helios” architecture following the OCP Open Rack Wide design, which is expected to facilitate quicker deployment and adaptability for demanding AI applications. HPE’s involvement extends to delivering this integrated architecture worldwide in 2026. In parallel, AMD and HPE are collaborating on the Herder supercomputer for the High performance computing (HPC) Center Stuttgart (HLRS) in Germany, built on the HPE Cray GX5000 platform. This system will deploy AMD Instinct MI430X GPUs and EPYC “Venice” processors to handle HPC and AI workloads at scale.
Lisa Su, Chair and CEO of AMD, said, “With ‘Helios’, we’re taking that collaboration further, bringing together the full stack of AMD compute technologies and HPE’s system innovation to deliver an open, rack-scale AI platform that drives new levels of efficiency, scalability, and breakthrough performance for our customers in the AI era.” Antonio Neri, president and CEO of HPE, said, “With the introduction of the new AMD ‘Helios’ and our purpose-built HPE scale-up networking solution, we are providing our cloud service provider customers with faster deployments, greater flexibility, and reduced risk in how they scale AI computing in their businesses.” Prof. Disaster Recovery (DR). Michael Resch, Director of HLRS, said, “Herder's system architecture will enable us to support both of these approaches, while also giving our users the ability to develop and benefit from new kinds of hybrid HPC/AI workflows.”
Herder is scheduled for delivery in the latter half of 2027 and aims to replace HLRS's existing flagship supercomputer. The expanded collaboration between AMD and HPE anticipates supporting large-scale AI infrastructure and HPC applications in coming years.