AMD stepped directly into Anthropic’s corner on July 22, pledging up to $5 billion and a wave of next-generation GPUs in a deal that gives the chipmaker its first equity stake in a frontier AI lab.
The partnership, unveiled at AMD’s Advancing AI event in San Francisco, centers on deploying 2 gigawatts of Instinct MI450 Series graphics processors inside AMD’s Helios rack systems starting next year. Each rack pairs the new MI455X chips with AMD’s freshly announced EPYC Venice processors, a 256-core Zen 6 server chip built on TSMC’s most advanced 2nm manufacturing process.
Lisa Su, AMD’s chief executive, described the arrangement as the full force of the company’s hardware lineup meeting one of AI’s fastest-growing customers. Beyond hardware, Anthropic will steer its Claude models toward improving AMD’s ROCm software stack, an area where Nvidia’s CUDA platform has long held the upper hand.
The investment arrives as Anthropic prepares for a potential public listing and as AI labs everywhere scramble to lock down computing capacity. Wall Street read the announcement as a sign that AMD is serious about breaking Nvidia’s grip on the AI chip market, though the revenue impact will not register until 2027 at the earliest.
