Anthropic Eyes Samsung for Custom AI Chip as AI Giants Flee Nvidia Dependency

Anthropic is in early-stage talks with Samsung Electronics to manufacture a custom AI chip on its 2-nanometer process, joining OpenAI, Google, and Amazon in a race to reduce reliance on Nvidia hardware.

Anthropic, the maker of Claude, is in early discussions with Samsung Electronics about manufacturing a custom AI chip, according to a report by The Information confirmed by multiple outlets. The project remains at an early stage — no chip design has been finalized, and Anthropic has not yet determined what the chip would be used for or how it would fit into a server.

The talks center on Samsung's 2-nanometer manufacturing process and its advanced chip-packaging facilities, which place processors and high-bandwidth memory closer together to boost data-transfer speeds. Samsung is one of three semiconductor companies — alongside SK hynix and Micron — that participated as strategic infrastructure partners in Anthropic's $65 billion Series H funding round, completed May 28.

A Chip Team Takes Shape

Anthropic hired Clive Chan in June, who was the second hardware engineer to join OpenAI's custom-chip program and worked on it from its earliest stages. The recruitment signals that Anthropic is building an internal team capable of designing specialized processors, even as the company stresses that "Nvidia GPUs, Google TPUs and AWS Trainium chips will continue to play a central role in our computing resources."

The move comes just one week after OpenAI unveiled Jalapeño, its first custom inference processor built with Broadcom — taken from initial design to production in nine months. Google has developed multiple generations of its Tensor Processing Units, while Amazon operates its Trainium line for AI training workloads.

Why Samsung

A manufacturing agreement would give Samsung a major AI customer as the South Korean chipmaker fights to challenge Taiwan Semiconductor Manufacturing Co. (TSMC) in advanced-node manufacturing. Samsung previously signed a $16.5 billion deal to manufacture next-generation AI chips for Tesla, and Google is reportedly considering Samsung for part of a future tensor processor.

South Korea on Monday unveiled a wider semiconductor investment plan under which Samsung and SK hynix are expected to invest approximately 800 trillion won ($523 billion) over the next decade — including four new fabrication plants and expanded high-bandwidth memory production.

The Bigger Picture

Anthropic's annualized revenue run rate reportedly surpassed $30 billion earlier this year, more than tripling from roughly $9 billion at the end of 2025. At that scale, the economics of custom silicon — lower per-inference costs, better energy efficiency, and infrastructure independence — become increasingly compelling.

The company is entering the custom-chip race later than its competitors, but rapidly rising demand for AI infrastructure is creating room for specialized alternatives to Nvidia's dominant GPUs.

Sources: UPI, TechCrunch