
AMD is adding Taalas’ model-in-silicon inference technology to its AI accelerator roadmap.
Anthropic plans to use AMD Instinct MI450 GPUs for Claude training and inference as AMD invests up to $5 billion, expanding AMD’s push into AI infrastructure.

AMD is investing up to $5 billion in Anthropic, while Anthropic plans to deploy up to 2 gigawatts of AMD Instinct MI450 GPUs for Claude. The GPUs will be used for both training and running Anthropic’s Claude models, making the agreement a major AI infrastructure commitment. The arrangement ties capital investment directly to demand for AMD’s AI hardware.
The deployment is planned around AMD Instinct MI450 GPUs in Helios server systems. The first gigawatt phase is expected to start in the first half of 2027. The systems pair MI455X GPUs with AMD EPYC “Venice” processors and AMD networking technology, and Anthropic already uses AMD’s MI355X GPUs.
AMD and Anthropic also plan a multi-year effort to use Claude to improve AMD’s ROCm software platform and GPU workloads. AMD will use Claude internally across its development teams. Anthropic co-founder Tom Brown says working with different hardware providers lets the company match workloads to the best-suited hardware.
For AMD, the partnership is another step in its effort to establish itself alongside Nvidia as a GPU supplier for AI companies. The Decoder notes that AMD has struck similar deals with Meta and OpenAI. Critics point to the circular nature of these agreements, where chip and cloud companies fund AI labs that then buy from them, while the question of whether AI labs can cover these bills independently remains open.

AMD is adding Taalas’ model-in-silicon inference technology to its AI accelerator roadmap.

A reported six-year cloud compute deal would add Norway-based AI infrastructure to Anthropic’s growing capacity push.

OpenAI and Anthropic’s autonomous AI hacks raise new questions under U.S. hacking law.

User demos show Opus 5 creating playable 3D prototypes with code-generated assets, physics, and music.