Anthropic Builds Custom AI Chip Team for Claude

According to Anthropic’s announcement on August 4, the company is forming an internal engineering team dedicated to developing custom chips designed specifically for its Claude AI models. The move confirms that the AI startup is exploring in-house silicon solutions to address the ongoing shortage of high-performance AI chips and support the growing demand for large-scale model training and inference.

Anthropic said it is recruiting engineers with expertise across both hardware and software development to work on the co-design of custom chips and AI models. The goal is to improve the speed, efficiency, and scalability of Claude deployments for customers running large AI workloads.

The company noted that custom silicon represents the latest step in its “multi-chip strategy.” Anthropic will continue to rely on a diversified hardware ecosystem, including technologies from partners such as Amazon Web Services, Google, NVIDIA, and AMD. However, Anthropic has not disclosed a timeline for its custom chip development or confirmed whether it plans to manufacture the chips independently.

Industry sources estimate that developing an advanced AI chip can cost around $500 million, as companies must not only build highly skilled engineering teams but also invest heavily in achieving reliable, high-quality semiconductor production.