AMD Begins Production of Helios, Claims 30% Advantage in Tokens per Dollar

AMD has entered the production phase of its AI data center rack-scale system, Helios. The company announced that Helios generates up to 30% more inference tokens per dollar compared to competing products. During the 'Advancing AI 2026' event on July 23, 2023, AMD unveiled Helios, stating that major AI companies are expected to adopt it at a gigawatt scale. Helios features a design that integrates 72 AMD Instinct MI455X GPUs, 18 6th Gen EPYC Venice CPUs, Pensando networking, and ROCm software, achieving 2.9 EF FP4 performance, 1.4 EF FP8 performance, 31TB of HBM4 memory, and 23.3TB/s memory bandwidth per GPU. AMD emphasized rack-level economics over single-chip performance, explaining that data center operators are considering costs related to power, memory, networking, and software. Helios is provided as a reference design for OEM and ODM companies to build systems, with mass deployment scheduled for the second half of 2026. The company plans to invest over $10 billion in the Taiwanese ecosystem, with Sanmina, Wiwynn, Wistron, and Inventec expected to support the mass production of Helios-based systems. AMD reported that its revenue for Q2 2026 is projected to be $11.5 billion, a 50% increase year-over-year, with data center revenue rising to $6.7 billion, up 107%.

This content is provided for general informational purposes only and doesn't constitute financial, investment, legal, or tax advice. Any events, rewards, online promotions, or related information mentioned herein should not be considered a recommendation, solicitation, or invitation to purchase, sell, trade, or otherwise deal in any crypto assets. Crypto assets are highly volatile and may result in loss. The availability of WEEX services, products, and related events may vary by region. You are responsible for ensuring that your participation is in accordance with applicable local laws and regulations.