Anthropic confirms it is building an in-house silicon team to design custom chips for Claude, co-designing hardware and models and using a "multi-chip approach" (Tom Carter/Business Insider)

Anthropic, the AI startup, has officially confirmed its plans to construct an in-house team for designing custom AI chips specifically for its Claude model, as reported by Business Insider. According to a job listing and an official statement, the company is actively recruiting engineers to contribute to the development of these chips, marking the first time Anthropic has publicly acknowledged this ambition.

The spokesperson for Anthropic elaborated on their strategy, stating that the company will engage in a "multi-chip approach," leveraging hardware from various leading tech firms such as AWS, Google, Nvidia, and AMD, in addition to their own bespoke silicon. This strategy underscores Anthropic's commitment to both efficiency and customization, aiming to enhance Claude's performance and align it with the specific requirements of their customers.

Additionally, the company has been exploring the potential of designing its own chips, with previous reports suggesting talks with Samsung Electronics as a possible manufacturing partner. The hiring of a "custom silicon team," highlighted in a recently posted job listing, reflects Anthropic's dedication to a hands-on approach in chip design and verification.

The role, offering a salary ranging from $320,000 to $485,000, is aimed at individuals with a proven track record in silicon finalization and shipping, emphasizing the importance of direct personal contribution and decisive action. As the demand for AI models surges, Anthropic's move to build its own chips is not only a strategic response but also a step towards reducing reliance on Nvidia, the current market leader in AI accelerator chips, and enabling more tailored computational solutions for their sophisticated models.

This development signals a broader trend among AI labs, with rivals like OpenAI, Meta, and Mistral also exploring their own silicon options, reflecting the intense competition and innovation in the field of AI hardware.