Anthropic Confirms It's Building Custom Chips for Claude

Anthropic confirmed it is building an in-house silicon team to design custom chips for its Claude AI model, according to a company spokesperson and a recently posted job listing. The first public acknowledgment of the plans confirms earlier reports that the startup was exploring the idea.

The spokesperson told Business Insider that Anthropic will co-design hardware and models, allowing Claude to run faster and more efficiently "at the scale our customers need." The company plans to adopt a "multi-chip approach," meaning hardware from AWS, Google, Nvidia, and AMD will remain central to Anthropic's scaling efforts alongside the custom designs.

The job listing, posted for a custom silicon engineer, seeks candidates with experience shipping semiconductor designs. The salary range is $320,000 to $485,000. The listing says the role requires "direct personal contribution" to finalizing and shipping designs and describes it as a position for someone who has "shipped silicon, has a realistic relationship with schedules, and is comfortable making consequential calls without a large organization behind them."

Anthropic joins OpenAI as the second major AI lab to develop its own chips. OpenAI unveiled a purpose-built chip codenamed Jalapeño, developed with Broadcom, in June. Both companies aim to reduce reliance on external suppliers and optimize performance for their specific model architectures.

Nvidia's Alpha Score is 76 out of 100, reflecting strong momentum. AMD's Alpha Score is 51, mixed. Both companies continue to dominate the training market, which remains the largest source of AI chip revenue for now.

The job listing confirms what Reuters and The Information reported earlier this year: that Anthropic was exploring chip design and held talks with Samsung about potential manufacturing. The company has now moved from exploration to execution. No timeline for first silicon has been disclosed.