Anthropic is reportedly exploring a partnership with Samsung to develop its own artificial intelligence chips, as AI companies increasingly seek greater control over the hardware powering their models.
According to a report by The Information, Anthropic has held discussions with Samsung about collaborating on a custom AI processor. The company has not yet finalized the chip’s intended purpose, its role within server infrastructure, or its performance specifications.
Company Evaluates Custom Chip Strategy
Anthropic declined to comment on the reported discussions with Samsung but said its computing strategy will continue to rely on a diversified hardware portfolio.
The company said processors supplied by Google, Amazon and Nvidia remain central to supporting its AI infrastructure, while providing no additional details regarding a potential collaboration with Samsung.
Developing proprietary chips has become an increasingly common strategy among AI companies seeking to optimize hardware for specific workloads while reducing dependence on third-party suppliers.
Industry Pushes Beyond Standard AI Hardware
The reported talks come as demand for AI computing resources continues to grow, placing pressure on the supply of advanced processors.
Nvidia remains the dominant supplier of AI chips, but several technology companies are investing in custom silicon designed to improve efficiency and better match the requirements of their AI systems.
Last week, OpenAI announced a partnership with Broadcom to develop its own inference processor, known as “Jalapeño,” which the company said offers improved performance per watt compared with competing products.
Google and Amazon have also expanded the use of proprietary tensor processing units (TPUs) across their cloud computing platforms to support AI workloads.
Samsung Expands AI Semiconductor Role
Samsung already plays a significant role in the AI semiconductor supply chain through its manufacturing partnership with Nvidia, producing chips used to train and operate AI models while also utilizing Nvidia’s software in its own chip manufacturing processes.
The South Korean company is also working with Nvidia on an AI chip manufacturing facility in South Korea and has previously discussed semiconductor collaboration with Google.
The reported discussions with Anthropic underscore the growing competition among AI developers to build customized computing platforms as demand for high-performance AI infrastructure continues to accelerate.