AMD announced a strategic investment of up to $5 billion in Anthropic and a commitment to supply up to 2 gigawatts of Instinct MI450 GPUs, with the first gigawatt arriving in H1 2027 in AMD Helios systems.

AMD will deploy up to 2 gigawatts of Instinct MI450 Series GPUs for Anthropic and invest up to $5 billion in equity, the companies announced jointly. The first gigawatt is scheduled to arrive in the first half of 2027, deployed in AMD Helios rack-scale systems.

The announcement marks AMD as both supplier and shareholder to Anthropic, a structure that contrasts with AMD's deal with OpenAI, where the customer received warrants for roughly 10% of AMD's stock. The joint press release also publicly confirmed for the first time that Anthropic already operates AMD's current-generation MI355X GPUs.

Anthropic's deployment will pair Instinct MI450X GPUs with EPYC "Venice" CPUs, Pensando networking, and ROCm software inside Helios racks. Each rack houses 72 accelerators, 31 terabytes of HBM4 memory, and up to 2.9 FP4 exaFLOPS of inference compute. The companies launched a multi-year engineering collaboration that will use Claude to optimize workloads for Instinct GPUs and accelerate ROCm development, while AMD integrates Claude across its engineering and product teams.

"Running across a diversified range of hardware lets us map the right workloads to the right hardware," said Tom Brown, co-founder and chief compute officer at Anthropic.

The partnership builds on existing usage; SemiAnalysis had previously reported that a YAML configuration file in an AMD senior director's public GitHub repository listed Anthropic as a customer with maximum priority status, equivalent to hyperscale customers like Meta. Prior to today's announcement, Anthropic's disclosed suppliers were Nvidia, Amazon, and Google.

Anthropic's first AMD gigawatt arrives behind other Helios deployments already committed. OpenAI's first gigawatt of AMD capacity is due in the second half of 2026, and Oracle is beginning deployment of 50,000 MI450 GPUs this quarter.

The 2-gigawatt commitment expands an already-diversified compute portfolio that includes Nvidia GPUs, Amazon's Trainium chips under Project Rainier, 3.5 gigawatts of Google TPU capacity supplied by Broadcom from 2027, and 300 megawatts of Nvidia compute rented from SpaceX's Colossus 1 data center—making AMD roughly the sixth distinct silicon source feeding Claude.