AMD shipped its Helios AI rack system to major customers including OpenAI, Meta, and Microsoft, directly challenging Nvidia's dominance in data center infrastructure.

The chipmaker unveiled the system at its Advancing AI conference in San Francisco on Thursday. CEO Lisa Su called Helios the industry's "highest-performance AI rack," designed to train and run frontier models at gigawatt scale.

Helios combines multiple processors into a single high-powered unit built for data centers. The system outperforms Nvidia's Vera Rubin rack across several metrics, according to The Register.

Anthropic and AMD announced a strategic partnership Wednesday to deploy up to two gigawatts of GPUs via Helios. Microsoft CEO Satya Nadella said Monday the company would expand Azure infrastructure with the new system.

Other confirmed customers include Oracle and Meta. The system was first revealed in 2025 and demonstrated at CES 2026 in January.

AMD's broader AI push

AMD also introduced its Venice-X CPU on Thursday, targeting data centers and high-computing workloads. The Venice-X launches in 2027 with 96 cores and 1,152 MB of 3D V-Cache.

Nvidia has historically dominated the AI infrastructure market with its Vera Rubin and Grace Blackwell rack systems. AMD's entry represents a significant challenge to that position.

The Helios system weighs as much as two compact cars and requires specialized data center infrastructure. AMD said leading AI companies plan deployments at gigawatt scale.

Su commented on the chip industry's trajectory during her remarks, though specific details were not disclosed. The company positions Helios as purpose-built for the "most demanding frontier models."

AMD expects to begin shipping Helios systems to customers later this year. The company has not disclosed pricing or specific performance benchmarks.