AMD and Anthropic sign up to $5B deal for AI computing capacity

In short: AMD says it will invest up to $5 billion in Anthropic and help power Anthropic’s AI systems with AMD chips.

What happened

AMD and Anthropic announced a partnership that includes AMD investing up to $5 billion in Anthropic. Anthropic is the company behind Claude, a popular AI assistant.

As part of the deal, Anthropic plans to use up to 2 gigawatts of AMD Instinct MI450 GPUs. A GPU is a type of computer chip that is often used to run and train AI models, since it can do many calculations quickly (like having a lot of tiny calculators working at once).

Anthropic will use these chips in AMD’s new Helios rack-scale system. Think of this as a packaged way to stack and connect large numbers of chips in data centers, which are big buildings full of computers.

The companies say they plan to deploy the first 1 gigawatt in the first half of 2027. This follows other recent Anthropic deals to secure data center capacity, including agreements mentioned with SpaceX and TeraWulf, plus earlier infrastructure partnerships involving Google, Broadcom, and Amazon.

AMD and Anthropic also say they will work together for multiple years on engineering. AMD plans to use Claude across its software development, engineering, and product work.

Why it matters

AI systems are expensive to run because they need a lot of computing power and electricity. Deals like this are basically about reserving space and equipment ahead of time, similar to booking a large block of seats before a busy event. For everyday users, more computing capacity can mean AI tools that respond faster, handle more people at once, and improve over time.

Source: The Verge AI