The initiative is in preliminary stages, and the company may ultimately opt to continue purchasing AI chips rather than designing them, as indicated by two people familiar with the situation and another person briefed on Anthropic’s plans. A source mentioned that the company has not yet committed to a specific design or assembled a team focused on this project.
A representative of the San Francisco-based company chose not to comment on the article.
Read More: Tesla is developing a new, smaller, cheaper EV, says report
The demand for its AI model Claude has surged in 2026, with the startup’s run-rate revenue now exceeding $30 billion, increasing from approximately $9 billion at the end of 2025, Anthropic reported earlier this week.
Anthropic utilizes various chips, including tensor processing units (TPUs) developed by Alphabet’s Google and Amazon’s custom chips to create and run its AI software and the Claude chatbot.
Earlier this week, Anthropic entered into a long-term agreement with Google and Broadcom, which assists in designing the TPUs. This agreement complements the company’s pledge to invest $50 billion in enhancing U.S. computing infrastructure.
Anthropic’s discussions reflect similar initiatives among major tech companies that are working on their own AI chip designs, including Meta and OpenAI.
Creating an advanced AI chip can require an investment of around half a billion dollars, according to industry experts, as companies must hire skilled engineers and ensure a defect-free manufacturing process.
