AMD invests $5B in Anthropic, committing up to 2GW of next-gen AI chips

AMD made a multibillion-dollar strategic investment in Anthropic, reported at $5 billion, alongside a commitment to supply up to 2 gigawatts of next-generation AI chips, with the first gigawatt slated to deploy in 2027. The deal is a significant compute-diversification move for Anthropic, reducing reliance on NVIDIA and Google TPUs and giving AMD a marquee frontier-lab customer for its data-center accelerators.

The investment lands amid a broader wave of compute-investment deals defining the week's industry mood — Microsoft-Mistral's European GPU buildout, Google's Frozen v2 chip, and Meta's Nvidia expansion all point to labs and chipmakers locking in multi-year, multi-gigawatt capacity. For AMD, taking an equity stake in a customer echoes NVIDIA's playbook and signals confidence in Anthropic's trajectory.

The timing is notable: Anthropic is reportedly in advanced IPO talks targeting October 2026, and shipped Claude Opus 5 the same week. Securing guaranteed compute at scale is essential to the capital narrative an IPO requires.

Skeptics will note that a 2027 first-gigawatt deployment is far off, and that circular financing — chipmakers investing in customers who then buy their chips — raises questions about demand durability. Watch for confirmation of the exact investment figure, IPO filing details, and whether the AMD partnership yields production silicon on schedule.