AMD and HPE have expanded their collaboration with the aim of accelerate the next generation of your open and scalable AI infrastructurebased on AMD computing technologies. Under the agreement, HPE will adopt the rack-scale AI architecture of AMD’s Helios platform, which will integrate a purpose-built HPE Juniper Networking switch in collaboration with Broadcom, as well as software for seamless, high-bandwidth connectivity over Ethernet.
Helios combines AMD EPYC CPUs, AMD Instinct GPUs, AMD Pensando advanced networking, and the AMD ROCm open software stack to deliver a cohesive platform optimized for the highest level of performance, efficiency, and scalability. The system is designed to simplify the deployment of large-scale AI clusters, enabling faster response times and more infrastructure flexibility in research, cloud, and enterprise environments.
This platform delivers up to 2.9 exaflops of FP4 performance per rack, with Instinct MI455X GPUs, next-generation AMD EPYC Venice CPUs, and AMD Pensando Vulcano NICs for expansion networks. All unified through the ROCm open software ecosystem, which provides flexibility for AI and HPC tasks.
Built on the OCP Open Rack Wide design, Helios accelerates deployment times and offers a scalable and flexible solution for AI tasks that require high computing power. This has allowed HPE to integrate differentiated technologies for its customers. Specifically, a scale-up Ethernet switch and software designed for Helios.
Developed as we have mentioned in collaboration with Broadcom, the switch is optimized to achieve high performance with AI tasks. To do this, it uses the Ultra Accelerator Link over Ethernet (UALoE) standard, which reinforces AMD’s commitment to open, standards-based technologies. HPE will offer AMD Helios rack-scale AI architecture in 2026.
Lisa Su, President and CEO of AMDrecalled that «HPE has been an exceptional, long-term partner to AMD, working with us to redefine what is possible in high-performance computing. With ‘Helios’, we take that collaboration even further, bringing together AMD’s end-to-end compute technologies solution and HPE’s systems innovation to deliver an open, rack-scale AI platform that drives new levels of efficiency, scalability, and breakthrough performance for our customers in the age of AI.«.
As to Antonio Neri, President and CEO of HPEhas highlighted that «during For more than a decade, HPE and AMD have pushed the boundaries of supercomputing, delivering multiple exascale-class systems and championing open standards that accelerate innovation. 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 lower risk in how they scale AI computing across their businesses.«.
On the other hand, Herder, a supercomputer developed on the HPE Cray Supercomputing GX5000 platformalso integrates AMD Instinct MI430X GPUs, as well as next-generation AMD EPYC Venice CPUs. It will be installed at the High Performance Computing Center Stuttgart (HLRS) in Germany, and is intended for large-scale HPC and AI tasks, both for scientific research and industrial innovation by European researchers and companies.
Herder is scheduled for delivery in the second half of 2027, and will go into operation at the end of 2027. It will replace the supercomputer currently in operation at the HLRS, known as Hunter.
