A job posting for the new team seeks engineers with experience bringing semiconductor designs through production, requiring candidates to demonstrate "direct personal contribution" to the final states of shipping semiconductor designs. The role reflects Anthropic's effort to establish an in-house hardware capability while coordinating chip development with its AI models.
Anthropic said it plans to "co-design" its hardware and models rather than treat them as separate development efforts. The company did not announce a manufacturing partner, although previous reports have suggested Samsung Electronics is being considered. Anthropic has not confirmed those reports.
The company indicated its custom silicon effort will focus on AI inference rather than model training. Inference hardware is used to run AI models after they have been trained, helping improve efficiency and performance for deployed services such as Claude.
The custom silicon initiative adds another layer to Anthropic's AI infrastructure strategy. Rather than relying solely on third-party hardware, the company plans to develop chips designed specifically for its own AI workloads.
For now, Anthropic said it will maintain its existing relationships with cloud infrastructure providers and chip vendors as it expands its internal hardware capabilities. The custom silicon program represents a new layer in the company's AI infrastructure strategy while preserving its current multi-vendor approach to compute.
This analysis is based on reporting from PC Mag.
Image courtesy of The Decoder.
This article was generated with AI assistance and reviewed for accuracy and quality.
