Anthropic is exploring the development of a custom AI chip and has held discussions with Samsung over a possible collaboration, according to The Information. The move places the startup among firms seeking greater control over AI hardware at a time of rising compute demand.
The development follows an earlier Reuters report that said Anthropic was considering in-house chip design to address concerns around AI hardware availability and supply constraints. The company has not disclosed detailed plans for the proposed processor.
According to the report, Anthropic has not finalized key aspects of the project, including the chip’s intended use, its integration into server systems, and its performance goals. The company has also declined to comment on discussions involving Samsung.
The Anthropic team stated that its current infrastructure strategy continues to rely on chips supplied by Google, Amazon, and NVIDIA. It added that it has no further updates on potential collaboration talks with Samsung.
The reported move reflects a wider trend among AI companies investing in custom silicon to improve efficiency and reduce reliance on dominant chip suppliers. NVIDIA continues to lead the market for AI accelerators, which power large-scale model training and inference workloads.
The shift comes as companies look for greater control over cost, performance, and supply chains as demand for AI computing expands across industries.
Anthropic’s reported discussions come shortly after other major technology firms moved further into custom chip development. OpenAI recently introduced its first custom inference chip, called ‘Jalapeno’, developed with Broadcom.
The company confirmed that the chip offers improved performance efficiency compared with existing alternatives. Amazon continues to scale its Trainium chips for AI training workloads, while Google is expanding access to its Tensor Processing Units, or TPUs, through its cloud infrastructure.
Samsung already plays a key role in the global semiconductor supply chain. The company manufactures advanced chips and works with major AI players, including NVIDIA, to support high-performance computing demand. It has also explored partnerships with other technology companies, including Google, as demand for specialised AI hardware continues to grow across the industry.
The discussions between Anthropic and Samsung highlight the increasing overlap between AI software developers and semiconductor manufacturers. This shows firms seek tighter integration between hardware and AI systems.