Key Takeaways
- Advanced Micro Devices has partnered with Celestica to create Helios, a revolutionary rack-scale AI platform
- Celestica is responsible for research, development, design, and production of scale-up networking switches within the platform
- These networking switches will link AMD’s upcoming Instinct MI450 Series GPUs for massive AI cluster deployments
- The Helios platform utilizes Open Compute Project Open-Rack-Wide architecture and Ultra Accelerator Link over Ethernet technology
- Customer availability is slated for late 2026; AMD shares gained approximately 1% in premarket trading Monday
Advanced Micro Devices and Celestica (CLS) have joined forces to launch an innovative rack-scale AI platform. Dubbed Helios, this system targets massive-scale AI training and inference applications spanning cloud computing, enterprise data centers, and research institutions.
Celestica’s responsibilities include handling the research and development, engineering design, and production operations for scale-up networking switches that form the backbone of the AMD Helios infrastructure. The switches leverage the Open Compute Project Open-Rack-Wide form-factor — an increasingly popular open-source standard in hyperscale computing facilities.
Advanced Micro Devices, Inc., AMD
The networking chips powering these switches are specifically designed to facilitate ultra-fast interconnectivity among AMD’s forthcoming Instinct MI450 Series GPUs. The platform employs Ultra Accelerator Link over Ethernet technology for scale-up connectivity, which is critical for maintaining rapid communication between GPU clusters.
With rack-scale AI infrastructure, the complete rack — rather than individual servers — functions as the primary computing unit. This approach integrates GPUs, high-bandwidth networking, and liquid cooling technologies into one unified system. The architecture aims to efficiently handle large language model training at unprecedented scales.
“Helios represents a new blueprint for AI infrastructure,” stated Forrest Norrod, AMD’s executive vice president and general manager of Data Center Solutions. He emphasized that it empowers customers to implement AI solutions with the performance, efficiency, and adaptability required for next-generation workloads.
Steven Dorwart, senior vice president at Celestica, noted that scaling AI infrastructure demands systems that can be deployed rapidly and reliably. Celestica brings its proven expertise in data center architecture, engineering capabilities, and supply chain management to the Helios initiative.
AMD’s Growing Market Position
AMD currently holds a market capitalization of approximately $315 billion and has delivered a 92% return over the trailing twelve months as AI infrastructure demand has surged. The Helios platform announcement arrives as AMD expands its presence in the data center GPU sector.
UBS maintains a Buy rating on AMD with a $310 price target, highlighting strong revenue growth opportunities extending through 2027. The investment bank has identified potential for AMD to secure a third major hyperscaler client for its data center offerings, with Microsoft mentioned as a probable partner.
Wolfe Research similarly rates AMD as Outperform, emphasizing the company’s server business momentum and its AI accelerator development pipeline as primary growth catalysts.
Additional Strategic Initiatives
In addition to Helios, AMD has recently finalized a multi-year licensing arrangement with Adeia Inc., providing access to Adeia’s semiconductor intellectual property portfolio while resolving all outstanding legal disputes between the organizations.
Avalon GloboCare has gained admission to AMD’s AI Developer Program, granting the company access to AMD’s development tools and resources for artificial intelligence projects.
AMD shares increased roughly 1% in premarket trading Monday after the Helios platform announcement. Celestica stock jumped approximately 3% during the same trading session.
AMD Helios is expected to reach customer availability in late 2026.





