AMD & HPE team up on open rack-scale AI data centre platform
AMD and HPE have announced an expanded collaboration to develop new open, rack-scale artificial intelligence solutions for data centres and research institutions. HPE plans to integrate the AMD "Helios" architecture into its next generation of AI platforms and will also provide advanced networking through a joint development with Broadcom.
Platform adoption
HPE is set to be one of the first original equipment manufacturers to adopt the AMD Helios architecture. The platform, engineered for large-scale AI workloads, brings together AMD's CPUs, GPUs, advanced networking, and open source software into a suite designed to address growing demand for AI compute infrastructure. HPE will make the Helios rack-scale AI architecture available globally in 2026.
Networking integration
Helios will incorporate a purpose-built HPE Juniper Networking switch, developed with Broadcom, to ensure high-bandwidth and low-latency connectivity for large AI clusters. The system leverages Ultra Accelerator Link over Ethernet, following industry open standards. This approach aims to support greater flexibility and efficiency for research, enterprise, and cloud customers by simplifying the deployment of large-scale AI clusters.
The Helios platform includes AMD's next-generation EPYC CPUs, Instinct MI455X GPUs, and Pensando Vulcano network interface cards. The architecture is built around the OCP Open Rack Wide design, allowing partners to streamline deployment timelines and provide scalable solutions for computationally intensive AI workloads.
Supercomputing project
As part of their collaboration, AMD and HPE are also supporting the development of "Herder," a new supercomputer commissioned by the High-Performance Computing Centre Stuttgart (HLRS) in Germany. Herder will be built on the HPE Cray GX5000 platform and use AMD's Instinct MI430X GPUs and next-generation EPYC CPUs. The system is expected to replace HLRS's current flagship and will begin service at the end of 2027.
Herder will provide computing resources for both traditional high-performance computing workloads and newer AI-driven applications. It is designed to enable European researchers and enterprises to advance scientific discovery and industrial innovation through sovereign, high-performance compute resources.
"The pairing of AMD Instinct MI430X GPUs and EPYC processors within HPE's GX5000 platform is a perfect solution for us at HLRS. Our scientific user community requires that we continue to support traditional applications of HPC for numerical simulation. At the same time, we are seeing growing interest in machine learning and artificial intelligence. 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. This platform will not only make it possible for our users to run larger, more powerful simulations that lead to exciting scientific discoveries, but also to develop more efficient computational methods that are only feasible with the capabilities that such next-generation hardware offers," said Prof. Dr. Michael Resch, Director, HLRS.
