Anthropic in Talks with Samsung for Custom AI Chip Manufacturing

Anthropic explores custom chip partnership with Samsung

Anthropic is holding early-stage discussions with Samsung Electronics about manufacturing a custom artificial intelligence processor, according to a report from The Information. The talks center on using Samsung's 2-nanometer manufacturing process and advanced packaging capabilities, though detailed design and testing work has not yet begun.

The Claude AI developer is evaluating the technical requirements for a specialized chip and how it would integrate into server infrastructure. Anthropic is simultaneously engaging with multiple chip-design firms and considering processors from Microsoft and British startup Fractile as it maps out its computing strategy.

Building internal hardware expertise

Anthropic's hardware ambitions became more concrete in June when it hired Clive Chan, who was the second hardware engineer to join OpenAI's custom chip program. Chan worked on that initiative from its inception before announcing his move to Anthropic on June 7. His recruitment signals that Anthropic is assembling an internal team capable of designing specialized processors as competition with OpenAI extends beyond AI models into hardware and data center infrastructure.

The company raised $65 billion in a Series H funding round completed May 28, reaching a post-investment valuation of $965 billion. Samsung Electronics, SK hynix, and Micron participated as strategic infrastructure partners, with Samsung being the only one among them that operates a major contract chip-manufacturing business.

Why it matters

Custom AI chips represent a strategic shift for major AI companies seeking to reduce dependence on Nvidia's dominant processors while controlling costs and improving energy efficiency. Anthropic's entry into this space, though later than competitors like Google and Amazon, comes as demand for specialized AI infrastructure accelerates. For Samsung, an Anthropic manufacturing contract would provide another marquee customer as it challenges Taiwan Semiconductor Manufacturing Co. in advanced processor production—particularly important given Samsung's previous yield challenges and its $16.5 billion Tesla chip deal.

Industry context and competitive landscape

Major technology companies are increasingly developing proprietary processors. Google operates multiple generations of tensor processing units, while Amazon Web Services runs Trainium chips for AI training. OpenAI and Broadcom unveiled Jalapeño, OpenAI's first custom inference processor, on June 24, with deployment expected by year-end.

Anthropic emphasized that custom chip development would complement rather than replace existing hardware relationships. The company stated that Nvidia GPUs, Google TPUs, and AWS Trainium chips will continue playing central roles in its computing resources.

Samsung's 2-nanometer process ranks among the most advanced semiconductor manufacturing technologies available. Smaller manufacturing nodes enable more transistors per chip, potentially boosting computing performance and energy efficiency. Advanced packaging techniques place processors, high-bandwidth memory, and other components closer together, increasing data-transfer speeds and reducing bottlenecks when running large AI models.

TrendForce projects that shipments of servers using cloud companies' custom application-specific integrated circuits will grow 44.6 percent in 2026, compared to 16.1 percent growth for servers using general-purpose graphics processors. South Korea announced a broader semiconductor investment plan on Monday 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.

These details were first reported by The Information and Asia Today.