Chipmaker AMD aims to rival Nvidia with its latest hardware release: a system of scale designed to power the computing needs of the world’s largest AI labs.
At the company’s sold-out Advancing AI conference in San Francisco on Thursday, AMD Chairman and CEO Dr. Lisa Su promoted the new AI rack system known as Helios – and its growing list of customers, including Microsoft – as the company prepares to ship it later this year. Su also laid out the company’s newest chips designed to feed the computing-hungry dragon that is the AI industry.
Rack systems combine multiple processors into a single high-power unit. They are designed for data centers, where they train and run AI models and other complex workloads.
Su called Helios the technology industry’s “highest performance AI rack,” adding that it was “designed to train and run the world’s most in-demand boundary models on a large scale.” The system will be deployed by leading AI companies on a gigawatt scale, the company said.
Nvidia has historically dominated this market with Vera Rubin and Grace Blackwell rack-scale systems. AMD is clearly looking to get in on the action. And Helios’ performance metrics seem to give it a real edge, beating Vera Rubin on a number of metrics, the Register reported.
Helios, first unveiled in 2025 and shown on stage in January at CES 2026, already has several well-known customers, including OpenAI, Meta, Oracle, Anthropic, and Microsoft, all of which have plans to use the system. Microsoft CEO Satya Nadella said Monday that the company will expand its Azure infrastructure with Helios. Meanwhile, Anthropic and AMD announced a strategic partnership on Wednesday to deploy two gigawatts of GPUs in a new rack system.
AMD also introduced on Thursday its Venice-X CPU, which is designed for data centers and to handle computing workloads. Venice-X is expected to be launched in 2027.
During his remarks, Su commented on the chip industry’s trajectory, saying that, by the year 2030, AI-powered chips will be a major part of the overall computer market. That’s because the industry is “seeing a step change in computing demand” driven largely by the rise of agent AI, he said.
“When you ask an agent to do something, it actually has a lot of steps, and it has to think, and it has to call tools, and it has to access data, and it has to keep doing it over and over again until it solves the problem, so you need a lot of GPUs to do all that,” the official said.
“Now we expect that by 2030, the AI acceleration market will reach about $1.4 trillion,” said Su. “That means, by the end of the decade, the AI accelerator market will approach the size of the entire semiconductor market today.”
“We expect that GPUs will make up a large portion of that market because the algorithms are still very young, and we continue to see the shift in workloads, and that allows for scalability in an all-silicon system,” he added.
If you shop through links in our articles, we may earn a small commission. This does not affect our editorial independence.
