AMD introduces Helios system to challenge NVIDIA dominance

01:57 / 24.07.2026·31·Technology

AMD, one of the world's leading semiconductor companies, has announced a new rack system called Helios designed for large AI laboratories and data centers. Unveiled at the Advancing AI conference in San Francisco, this device aims to directly challenge NVIDIA's market dominance. This is reported by Techcrunch.com .

According to AMD CEO Lisa Su, Helios is currently the highest-performance AI system in the technology industry. This device is specifically designed to train and run the world's most complex and demanding AI models. The system integrates multiple processors into a single, ultra-powerful block, which is critical for projects requiring gigawatt-scale computing power.

Competition and strategic partnership

Until now, NVIDIA has held absolute leadership in this market with its Vera Rubin and Grace Blackwell systems. However, according to The Register, the performance metrics of the Helios system surpass the Vera Rubin system in several parameters. This situation opens new opportunities for AMD in the AI infrastructure market.

Interest in the new system is already high. AMD has already established partnerships with tech giants such as OpenAI, Meta, Oracle, and Anthropic. In particular, Microsoft CEO Satya Nadella announced that the company will expand its Azure cloud infrastructure using Helios systems. Anthropic has announced a strategic partnership with AMD to deploy up to two gigawatts of GPU capacity.

Future forecasts and new chips

During the conference, AMD did not limit itself to Helios. The company also showcased the new Venice-X CPU designed for data centers and high-performance computing workloads. This new chip is expected to hit the market in 2027. Speaking about the future of the AI chip market, Lisa Su emphasized that the market size in this sector is expected to reach $1.4 trillion by 2030.

According to the AMD executive, the onset of the agentic AI era will dramatically increase the demand for computing power. Such systems perform multi-step logical processes to solve problems, accessing databases repeatedly. This, in turn, requires thousands of powerful GPUs. By the end of the decade, the AI accelerator market is expected to equal the size of the entire semiconductor market today.