Anthropic explores designing its own AI chips as demand surges
Anthropic is exploring whether to design its own artificial intelligence chips, three sources said, as the company responds to shortages of processors needed to build and run advanced AI systems.
The effort remains in early stages. The company has not committed to a specific design, assembled a dedicated team, or decided it will proceed rather than continue buying chips from existing suppliers, according to two people with knowledge of the discussions and one person briefed on the plans.
An Anthropic spokesperson declined to comment.
Revenue acceleration drives infrastructure needs
Anthropic's revenue has accelerated sharply this year. The company's run-rate revenue now exceeds $30 billion, up from roughly $9 billion at the end of 2025, the company said earlier this week.
The San Francisco-based startup currently uses chips from multiple suppliers, including tensor processing units designed by Google and processors from Amazon, to develop and operate its Claude chatbot.
Earlier this week, Anthropic signed a long-term agreement with Google and Broadcom, which helps design Google's TPUs. The deal supports the company's commitment to invest $50 billion in U.S. computing infrastructure.
Industry pattern mirrors Meta and OpenAI efforts
Anthropic's exploration mirrors similar work at other large technology companies. Meta and OpenAI are also designing their own AI chips to reduce reliance on external suppliers and manage costs.
Developing an advanced AI chip costs roughly $500 million, according to industry sources. Companies must hire specialized engineers and invest in manufacturing processes to prevent defects.
For professionals building AI systems, understanding chip architecture and infrastructure constraints has become central to development strategy. Learn more about generative AI and LLM foundations and AI for IT and development to understand how hardware decisions shape system design.
Your membership also unlocks:
