AMD Challenges Nvidia's AI Dominance with Helios Rack System
AMD is stepping up its competition with Nvidia by unveiling Helios, a rack-scale AI system designed to power the world's largest AI labs. At the sold-out Advancing AI conference in San Francisco, CEO Dr. Lisa Su showcases the system as the tech industry's highest-performance AI rack, built to train and run demanding frontier models at massive scale. The system already secures a roster of major customers, including Microsoft, OpenAI, Meta, Oracle, and Anthropic, with deployments planned at gigawatt-scale.
Helios appears to give AMD a genuine shot at challenging Nvidia's dominance in the rack-scale market, where Nvidia's Vera Rubin and Grace Blackwell systems currently lead. According to reports, Helios beats Vera Rubin across several performance metrics. The momentum is building as Microsoft announces plans to expand Azure infrastructure with Helios, while Anthropic partners strategically with AMD to deploy up to two gigawatts of GPUs through the new system. AMD also introduces its Venice-X CPU for data centers, expected to launch in 2027.
Su emphasizes the broader trajectory of the chip industry, predicting that AI chips will become a massive portion of the overall computing market by 2030. She points to the rise of agentic AI as a key driver of this step change in compute demand. Because AI agents must reason, call tools, access data, and iterate repeatedly to solve problems, the need for GPU power continues to scale dramatically, reinforcing AMD's push to capture a larger share of this rapidly expanding market.