AMD's Helios rack system beats Nvidia Vera Rubin in key benchmarks

AMD officially unveiled Helios, its rack-scale AI system aimed directly at Nvidia's dominant Vera Rubin and Grace Blackwell platforms. At the Advancing AI conference in San Francisco, CEO Lisa Su touted Helios as the industry's highest-performance AI rack system, built for gigawatt-scale data centers to train and run the most demanding frontier models. According to reports, Helios beats Nvidia's Vera Rubin on multiple performance metrics. The system already has a strong customer base, including Microsoft, OpenAI, Meta, Oracle, and Anthropic. Microsoft CEO Satya Nadella confirmed Azure infrastructure expansion with Helios, while Anthropic and AMD announced a strategic partnership to deploy up to two gigawatts of GPUs via the system. Additionally, AMD introduced the Venice-X CPU, designed for high-computing data center workloads, expected to launch in 2027.

During her keynote, Su highlighted a massive shift in compute demand driven by agentic AI, where each task requires dozens of reasoning steps, tool calls, and data accesses—all needing vast GPU resources. She projected the AI accelerator market will reach $1.4 trillion by 2030, nearly matching the size of the entire semiconductor market today. Su emphasized that GPUs will dominate due to the evolving, programmable nature of AI workloads. AMD's Helios represents a serious challenge to Nvidia's near-monopoly in AI infrastructure, offering customers a competitive alternative for large-scale training and inference. With major commitments from hyperscalers and AI labs, Helios could reshape the data center landscape and accelerate the deployment of next-generation AI systems.