Anthropic said it is building an in-house chip design team to develop custom processors for its Claude family of artificial intelligence models, marking a significant step in the company’s effort to secure computing power as demand for AI continues to surge.
The San Francisco-based startup has begun recruiting engineers with expertise in hardware and software co-design to build chips optimized specifically for Claude. The company said the new effort aims to improve the performance, efficiency and scalability of its AI systems.
The move comes as AI developers race to secure access to advanced semiconductors, which remain in short supply because of the rapid expansion of generative AI services.
READ: Anthropic signs $19 billion AI data center lease with TeraWulf (July 6, 2026)
Despite launching its chip initiative, Anthropic said it will continue pursuing a “multi-chip” strategy. The company plans to keep using processors from Amazon Web Services, Google, Nvidia and AMD while its custom silicon program develops.
Anthropic did not disclose when its first in-house chip could be ready or whether it intends to manufacture the processors itself. Industry analysts estimate that developing a cutting-edge AI chip can cost roughly $500 million before manufacturing expenses are included.
The announcement formally confirms a strategy that Reuters reported earlier this year, when sources said Anthropic was exploring the development of proprietary AI chips to reduce its reliance on outside suppliers.
READ: Alibaba bans use of Anthropic’s Claude Code over alleged security risks (July 3, 2026)
The company joins a growing list of major AI developers investing in custom silicon. OpenAI, Meta and several other technology companies are pursuing proprietary chips as they seek to lower infrastructure costs, improve performance and reduce dependence on Nvidia, whose graphics processors have become the industry’s dominant hardware for training and running AI models.
Anthropic’s decision reflects the broader industry trend toward designing AI hardware alongside AI models, allowing companies to optimize both together while maintaining access to multiple chip suppliers during a period of intense competition for computing resources.