Anthropic building in-house AI chip design team

AI startup Anthropic confirms it is building an in-house team to design custom chips for its Claude AI model, responding to surging user demand. The goal is to co-design hardware and models for greater speed and efficiency. This move follows a trend set by rivals like OpenAI and Meta to reduce reliance on market-dominant Nvidia, though Anthropic will maintain its multi-chip strategy with existing partners.

Anthropic in-house chip team

• ▪The decision to build a chip team is a response to surging demand for Anthropic's AI models.
• ▪The company's goal is to co-design hardware and models to make its Claude AI model run faster and more efficiently.
• ▪Anthropic has reportedly held discussions with Samsung Electronics as a potential manufacturing partner for its custom chips.
• ▪Anthropic is building an in-house team to design its own custom AI chips for its Claude AI model.

Multi-chip hardware strategy

• ▪The company has existing deals with AWS, Google, Nvidia, and AMD to access AI computing hardware.
• ▪Anthropic will continue to use a "multi-chip approach," with hardware from partners remaining central to its scaling efforts.

AI industry chip design trend

• ▪In June, OpenAI unveiled its "Jalapeño" chip, which was developed with Broadcom.
• ▪Other major AI labs designing their own chips include OpenAI, Google DeepMind, and Meta.
• ▪Designing custom AI chips allows companies to reduce reliance on Nvidia, which dominates the AI accelerator chip market.

Engineering hiring requirements

• ▪Job candidates must demonstrate "direct personal contribution" to the finalization and shipping of semiconductor designs.
• ▪A listed engineer position for the team has a salary range of $320,000 to $485,000.
• ▪Anthropic is hiring engineers for a "custom silicon team" with expertise in chip design and verification.

2 sources