Amid AI Chip Shortages, Anthropic Considers Developing Its Own AI Chips to Boost Computing Control

Against the backdrop of AI chip shortages, Anthropic is evaluating the feasibility of developing its own AI chips to enhance computing power control, joining tech giants like Meta and OpenAI in this move. The plan is not yet in a substantive stage; the company may still rely on external chip supplies, with no clear design direction or dedicated team established.

Anthropic currently uses chips from Google, Amazon and other companies for model training and inference. Recently, it signed long-term cooperation agreements with Google and Broadcom to deepen TPU layout, alongside its 50 billion USD computing infrastructure investment plan. As Claude's demand surged in 2026, the company's annualized revenue exceeded 30 billion USD, a significant increase from 9 billion USD at the end of 2025. Industry insiders note that developing advanced AI chips costs about 500 million USD, involving high-end team building and complex manufacturing process control.