Helios: AMD's New System to Challenge NVIDIA in the AI Market

AMD Enters the AI Battle with a New Weapon

In an ambitious move to break NVIDIA's dominance over the AI infrastructure market, AMD has unveiled its new system "Helios" — a fully integrated rack-scale computing platform designed to meet the demands of the world's largest AI laboratories. The announcement came during the sold-out "Advancing AI" conference in San Francisco, chaired by AMD CEO Dr. Lisa Su.

What Sets the Helios System Apart?

AMD describes Helios as the highest-performing rack-scale system in the AI sector, purpose-built to train and run the most complex frontier AI models at massive scale. The company plans to deploy the system at leading AI firms at power consumption levels reaching the gigawatt range.

Rack-scale computing systems consolidate a large number of processors into a single high-capacity unit, intended for data centers that handle AI model training and compute-intensive applications. According to specialized reports, Helios outperforms NVIDIA's "Vera Rubin" system on several benchmarks.

Major Customers Join Helios

AMD has assembled a roster of prominent names from the technology and AI world for its new system, including:

• Microsoft: CEO Satya Nadella announced the expansion of Azure's infrastructure using Helios.
• Anthropic: Entered into a strategic partnership with AMD to deploy up to two gigawatts of GPU capacity through the new system.
• OpenAI, Meta, and Oracle: All announced plans to incorporate Helios into their infrastructure.

This broad coalition gives AMD strong commercial credibility in its challenge against NVIDIA, which has long dominated the market with its "Vera Rubin" and "Grace Blackwell" platforms.

Venice-X Processor: Investing in the Future of Data Centers

Alongside Helios, AMD announced its new "Venice-X" data center processor, designed to handle intensive compute workloads. The processor is expected to launch commercially in 2027, complementing AMD's comprehensive portfolio serving the AI sector.

Lisa Su's Vision for the Future of the Chip Market

During her keynote, Lisa Su painted a striking picture of the chip industry's future, grounded in the accelerating demand generated by generative AI applications and agentic AI. She explained that assigning a single task to an AI agent can require dozens of sequential computational steps — reasoning, tool calls, and data retrieval — which accounts for the ever-growing need for GPUs.

Su projected that the AI accelerator market will reach approximately $1.4 trillion by 2030, roughly equivalent to the size of the entire semiconductor market today. She also affirmed that GPUs will continue to dominate this market due to their software programmability, which makes them best suited to adapt to continuously evolving algorithms.

AMD vs. NVIDIA: Is the Balance of Power Shifting?

NVIDIA still leads this market by a clear margin, yet AMD is today demonstrating genuine competitive intent — particularly with its diverse customer base and competitive technical performance. The question remains open: is Helios alone enough to change the rules of the game, or does AMD still have a long road ahead in breaking NVIDIA's entrenched dominance? What is certain is that the coming months will be decisive as the system begins shipping and operating in the field.

✦ بقلم فريق دروب أيديا