AMD and Celestica announced a strategic collaboration to bring the open standards-based “Helios” rack-scale AI platform to market. The collaboration combines AMD's capabilities in high-performance and AI computing with Celestica’s expertise in advanced networking switch technologies. Celestica will handle the R&D, design, and manufacturing of scale-up networking switches for the AMD “Helios” rack-scale AI architecture. The architecture follows the Open Compute Project (OCP) Open-Rack-Wide (ORW) form-factor.
The scale-up switches will use advanced networking silicon to support high-speed interconnects for the next-generation AMD Instinct MI450 Series GPUs. This enables computing optimized for large-scale AI clusters. The networking switches will employ the Ultra Accelerator Link over Ethernet (UALoE) architecture for scale-up connectivity, consistent with the open standards-based design of the “Helios” platform. The AMD “Helios” platform will become available to customers in late 2026.
The companies are working together to support deployments of “Helios” in cloud, enterprise, and research environments. The effort addresses the need for solutions that reduce time-to-value and enhance supply chain resiliency for organizations investing in AI infrastructure. Steven Dorwart, SVP and GM, Hyperscalers at Celestica, stated that deploying AI at scale requires infrastructure delivered quickly, consistently, and with expected performance. The collaboration integrates Celestica’s global engineering, manufacturing, and supply chain capabilities with AMD’s innovation in high-performance computing.
Forrest Norrod, executive vice president and general manager of the Data Center Solutions Business Group at AMD, described “Helios” as a new blueprint for AI infrastructure that provides performance, efficiency, and flexibility for next-generation workloads. He noted the partnership leverages Celestica’s expertise in networking switch technology alongside AMD’s leadership in high-performance and AI computing.
AMD drives innovation in high-performance and AI computing with a portfolio that includes AI-optimized CPUs, GPUs, networking, and software for full-stack AI solutions
