AMD bets $5 billion on Anthropic’s computing future

The collaboration hinges on the rollout of AMD’s Helios rack-scale systems, with the first gigawatt scheduled for deployment by early 2027. This infrastructure expansion complements Anthropic’s existing arrangements with players like Google, Amazon, and Broadcom. By integrating its hardware directly into Anthropic’s pipeline, AMD aims to stabilize its position as a viable alternative to the Nvidia-dominated AI hardware market.

Beyond the raw processing power, the companies have committed to a multi-year engineering pact. AMD will embed Anthropic’s Claude models across its internal software development and engineering product stacks. Tom Brown, Anthropic’s cofounder and chief compute officer, noted that the partnership is designed to secure necessary capacity while optimizing the infrastructure specifically for training and serving Claude. This move follows a wave of aggressive data center acquisitions by Anthropic, including recent high-profile deals with SpaceX and TeraWulf, as the startup races to scale its compute capabilities.