Anthropic Starts Building a Custom Chip Team for Claude

Anthropic is creating an internal semiconductor-design team as the developer of Claude seeks greater control over the computing infrastructure behind its artificial-intelligence models. The company said it is hiring engineers with experience across hardware and software to co-design custom chips and AI models. The objective is to make Claude operate faster and more efficiently at the scale required by enterprise customers. Anthropic’s decision reflects the growing cost and scarcity of advanced AI computing. Frontier-model developers require enormous quantities of accelerators for training and inference. Designing hardware around a company’s specific model architecture can potentially improve performance, reduce energy use and lower long-term operating costs. The project does not mean Anthropic is immediately replacing its current suppliers. The company said custom silicon will become part of a broader multi-chip strategy that continues to use infrastructure and technology from Amazon Web Services, Google, Nvidia and AMD. Anthropic did not provide a production timeline or say whether it will manufacture chips itself. The move places Anthropic more directly into the infrastructure competition surrounding AI models. Model capability remains important, but access to processors, memory, networking, power and manufacturing capacity increasingly determines how quickly an AI company can expand.