Anthropic moves to build custom chips for Claude, following OpenAI and Google's lead

Anthropic has confirmed it is building a dedicated team to design custom silicon chips that will power its Claude models, joining a growing list of major AI labs pursuing hardware independence.

The move came to light after Business Insider spotted job postings for senior semiconductor engineers on Anthropic's career page. An Anthropic spokesperson subsequently verified the initiative to both Business Insider and TechCrunch, adding that the company plans to pursue a diversified hardware strategy rather than relying exclusively on proprietary chips.

The Broader Silicon Race#

Anthropic's custom chip efforts are part of a wave reshaping the AI infrastructure landscape. OpenAI recently unveiled a chip called Jalapeño designed specifically for large-language-model inference, developed in partnership with Broadcom. Meanwhile, Google has long run its models on custom silicon, while Meta and Mistral are similarly investing in bespoke hardware solutions.

The strategic rationale is straightforward: tailor-made chips can improve efficiency and reduce costs compared to general-purpose processors, giving AI firms greater control over their computational pipelines as they scale.

Mixed Supply Strategy#

Anthropic's approach differs slightly from a full pivot to proprietary hardware. According to the company's statement, it will maintain a "multi-chip approach"—continuing to incorporate third-party semiconductors alongside its own designs. Reports have suggested Samsung could serve as a manufacturing partner for Anthropic's silicon.

This hedged strategy reflects the reality that chip design and manufacturing remain complex, specialized endeavors; balancing custom silicon with off-the-shelf options allows Anthropic flexibility as it grows.