Anthropic Confirms Internal Chip Team for Custom Claude AI Processors

AI company Anthropic has confirmed it is forming an internal semiconductor team to design custom chips for its Claude models. According to the company's job postings and official statements, the team is recruiting engineers to co-design hardware and AI models, enabling Claude to run faster and more efficiently at a scale that meets customer demand. This marks the first public confirmation of the initiative by Anthropic. A company spokesperson also said Anthropic will continue to pursue a "multi-chip strategy," with hardware from Amazon Web Services, Google, Nvidia and AMD remaining core components of its plan to expand compute capacity.

Developing AI chips in-house is costly and technically demanding, but it can help reduce dependence on Nvidia, which currently dominates the AI accelerator chip market. In-house chips also allow AI labs to tailor their compute architecture to their model requirements, improving compute utilisation efficiency.

Anthropic is not the first AI company to enter chip design. In June 2026, rival OpenAI unveiled a dedicated chip, codenamed "Jalapeño," developed jointly with Broadcom. The chief executive of French AI startup Mistral has also said the company is considering developing its own chips. In addition, cloud providers including Google, Microsoft, Amazon and Meta are developing custom chips, with Google's in-house Tensor Processing Unit considered the most promising challenger to Nvidia's dominance.