Artificial intelligence lab Anthropic is currently evaluating the possibility of creating its own AI chips, as it and other industry players face a shortage of specialized processors required to develop and power next-generation AI systems. These plans remain in the preliminary phase, and the company may ultimately choose to continue purchasing chips rather than designing them, insiders familiar with the matter.
At this stage, Anthropic has not committed to a particular chip design nor assembled a dedicated engineering team to pursue the project. A spokesperson for the San Francisco-based firm declined to provide comments on these developments.
Demand for Anthropic’s AI model, Claude, has surged in 2026, with the startup’s run-rate revenue exceeding $30 billion, a significant increase from approximately $9 billion at the close of 2025. To develop and operate Claude, Anthropic currently utilizes a variety of chips, including tensor processing units (TPUs) engineered by Alphabet’s Google, as well as processors from Amazon.
In a notable advancement, Anthropic recently secured a long-term agreement with Google and Broadcom, a key designer of TPUs. This partnership aligns with Anthropic’s broader commitment to invest $50 billion in enhancing U.S. computing infrastructure.
Anthropic’s exploration of in-house chip design reflects a broader industry trend, with major technology companies such as Meta and OpenAI also pursuing the development of proprietary AI chips. Creating a sophisticated AI processor is a costly endeavor, often requiring investments nearing half a billion dollars. This includes hiring highly skilled engineers and ensuring flawless manufacturing processes to avoid defects.
