Anthropic to deploy 2 gigawatts of AMD GPUs in $5 billion（約8000億円） deal

AMD announced it will invest up to $5 billion（約8000億円） in Anthropic as part of a hardware deployment agreement that will see Anthropic deploy up to 2 gigawatts of AMD Instinct MI450 GPUs in Helios server systems. These systems pair MI455X GPUs with AMD EPYC "Venice" processors and AMD networking technology to train and run Anthropic's Claude models. The first gigawatt phase is scheduled to begin in the first half of 2027. Anthropic already uses AMD's MI355X GPUs, and this agreement expands that relationship significantly.

Beyond the hardware deployment, both companies have committed to a multi-year effort to use Claude to improve AMD's ROCm software platform and GPU workloads. AMD will deploy Claude internally across its development teams, leveraging the AI system to optimize its own software and hardware offerings. Tom Brown, Anthropic's co-founder, stated that working with different hardware providers allows the company to match workloads to the best-suited hardware, reflecting a strategy of vendor diversification rather than single-supplier dependence.

For AMD, the deal represents a major competitive move to establish itself alongside Nvidia as a GPU supplier for AI companies. The company has pursued a similar strategy with Meta and OpenAI, using capital investment to secure hardware deployment commitments from leading AI labs. However, critics have noted the circular nature of these agreements: chip makers and cloud companies provide funding to AI laboratories, which then use that capital to purchase hardware from the same funders. This pattern raises an unresolved question about whether AI companies will eventually be able to independently cover their substantial infrastructure costs.

AMD's $5 billion（約8000億円） investment in Anthropic marks a significant effort to establish itself as a primary GPU supplier for leading AI companies, following similar strategic deals with Meta and OpenAI. The deployment of up to 2 gigawatts of AMD hardware for Claude represents a substantial commitment that positions AMD directly against Nvidia in the competitive AI infrastructure market. By front-loading capital investment, AMD gains assurance that a major AI lab will purchase its latest hardware at scale.

The deal structure reveals a pattern the article itself acknowledges: chip makers and cloud providers fund AI laboratories, which then deploy that capital by purchasing hardware from those same funders. This circular relationship raises an important unanswered question: whether AI companies will eventually generate sufficient independent revenue to cover their infrastructure costs without ongoing external investment. For now, Anthropic's commitment to deploy AMD hardware—and to collaborate on software optimization—offers AMD both immediate revenue certainty and the ability to improve its competitive software ecosystem through real-world Claude workloads.

No comments yet. Be the first to share your thoughts!