According to multiple sources, artificial intelligence startup Anthropic is exploring the possibility of designing its own chips. The company and its rivals are grappling with a shortage of artificial intelligence chips that are critical for powering and developing more advanced artificial intelligence systems.
According to two people familiar with the matter and one person familiar with Anthropic's plans, the plan is in its early stages and the company may ultimately choose to buy artificial intelligence chips rather than design its own. One of the sources said the company has yet to finalize a specific design and has not assembled a dedicated team to work on the project.
Anthropic said earlier this week that demand for its artificial intelligence model Claude is accelerating in 2026, and the company expects annualized revenue to exceed $30 billion this year, up from about $9 billion at the end of 2025.
Anthropic uses a variety of chips, including tensor processing units (TPUs) designed by Alphabet's Google and chips from Amazon, to develop and run its artificial intelligence software and chatbot Claude.
Earlier this week, Anthropic signed a long-term agreement with Google and Broadcom, which will help design Google's TPU. The agreement builds on Anthropic's commitment to invest $50 billion to strengthen U.S. computing infrastructure.
Anthropic’s talks echo similar efforts underway by big tech companies like Meta and OpenAI, which are looking to design their own AI chips.
According to industry sources, designing an advanced AI chip can cost around $500 million, as companies need to hire skilled engineers and invest money to ensure there are no defects in the manufacturing process.
