Anthropic reportedly opens early talks with Samsung to make a custom AI chip on 2nm process
Anthropic has reportedly held early talks with Samsung Electronics about manufacturing a custom AI chip on Samsung's 2-nanometer process, per reporting from July 2.
Anthropic has reportedly started early work on a custom AI chip and held talks with Samsung Electronics as a potential manufacturing partner, using Samsung’s 2-nanometer process and advanced packaging. The discussions are early, and the project has no finalized design, purpose or server-integration plan yet.
The move would put Anthropic on the same path as rivals building in-house silicon to cut their dependence on off-the-shelf accelerators. Designing a custom chip is a costly, multi-year bet that only pays off at large scale, but it can lower the per-query cost of running models and reduce exposure to a supply chain dominated by Nvidia.
Anthropic recently hired Clive Chan, an early member of OpenAI’s custom-chip team, as part of an engineering buildout, according to the reporting. Samsung is already an investor, having taken part in Anthropic’s $65 billion funding round in May 2026, which would give a manufacturing deal an existing relationship to build on.
Anthropic has not confirmed a chip program and said it will keep relying on a diversified hardware stack, including chips from Google, Amazon’s Trainium and Nvidia, and is in discussions with Microsoft and UK startup Fractile for additional supply. That hedging matters: an early-stage design talk is not a committed tape-out, and most such efforts take years to reach silicon, if they ship at all.
The move follows OpenAI’s disclosure of its own custom inference chip, Jalapeno, built with Broadcom. Whether Anthropic’s talks with Samsung firm into a signed manufacturing deal, or stay exploratory, is the next marker.
Founder and Chief Editor of Data Phoenix — a San Francisco Bay Area media and education platform focused on AI and Data.
More news

AWS releases six open-source Hugging Face deployment skills for SageMaker

Google Research releases MilleMiglia logistics benchmark generator

AWS launches AgentCore Runtime V2 with elastic memory and snapshot starts
