AMD unveils Helios AI rack-scale system to compete with Nvidia’s Vera Rubin

AMD announced Helios, a rack-scale AI system designed to train and run the most demanding frontier models at massive scale, positioning it as a competitor to Nvidia’s Vera Rubin and Grace Blackwell systems. The company stated Helios is built for data center deployments by leading AI companies at gigawatt-scale. Shipments are scheduled to begin later this year.

AMD Chair and CEO Dr. Lisa Su highlighted Helios as the “highest-performance AI rack” during the company’s Advancing AI conference, emphasizing its capability to handle compute-intensive workloads. Su also introduced the Venice-X CPU, a data center chip slated for a 2027 launch, designed to support high-computing workloads.

Helios has already secured commitments from major AI labs, including Microsoft, OpenAI, Meta, Oracle, and Anthropic. Microsoft’s Satya Nadella separately announced plans to expand Azure infrastructure with Helios, while Anthropic and AMD formalized a strategic partnership to deploy up to two gigawatts of GPUs via the new system.

AMD’s announcement follows reporting from The Register that Helios outperforms Nvidia’s Vera Rubin across several metrics. Nvidia currently leads the rack-scale AI hardware market with its Vera Rubin and Grace Blackwell systems, making AMD’s entry a notable shift in competitive dynamics.

During the conference, Su projected that AI accelerators could represent a $1.4 trillion market by 2030, approaching the size of the entire semiconductor market today. She attributed this growth to rising demand driven by agentic AI workloads, which require repeated tool calls, reasoning, and data access, necessitating large-scale GPU deployment.