Anthropic is in talks with Samsung to develop an AI chip
Anthropic is reportedly exploring a custom AI chip with Samsung as leading AI firms look to reduce reliance on Nvidia and secure more computing power.
The AI chip race is getting more intense. Anthropic is reportedly in discussions with Samsung about developing a new custom AI chip, a move that could strengthen the company’s long-term computing strategy.
The talks, reported by The Information and covered by TechCrunch, suggest that Anthropic is becoming more serious about building hardware tailored to its own artificial intelligence needs. Here's everything you need to know!
Why Anthropic is exploring custom chips
Anthropic, the company behind the Claude family of AI models, needs vast amounts of computing power to train and run its systems. As demand for advanced AI grows, access to high-performance chips has become one of the biggest challenges for AI companies.
Custom chips can help companies design hardware for specific tasks. In simple terms, instead of relying only on general-purpose processors, an AI company can create a chip that is better suited to running or training its models. This can improve efficiency, reduce costs over time and give companies more control over their infrastructure.
The report said Anthropic has not finalised the chip's intended use, server integration, or expected performance. This means the discussions appear to be at an exploratory stage rather than a confirmed product plan.
Samsung’s role in the AI supply chain
Samsung is already a major player in the global semiconductor industry and has deep ties to the AI hardware market. The company works with Nvidia and produces advanced chips for training and running AI models. Samsung also uses Nvidia software in its chip manufacturing processes.
A partnership with Anthropic would give Samsung another opportunity to expand its position in AI infrastructure. For Anthropic, Samsung could offer manufacturing expertise and chip development capabilities at a time when competition for AI hardware remains high.
The Nvidia factor
Nvidia remains the dominant supplier of AI chips, and its graphics processing units are widely used by companies building large AI models. However, many AI firms are now trying to reduce their dependence on a single supplier.
Anthropic has said that a diversified hardware stack, including chips from Google, Amazon and Nvidia, remains central to its compute strategy. A potential Samsung collaboration would fit into that broader approach, giving the company more flexibility as its computing needs increase.
How rivals are moving
Anthropic is not alone in exploring custom hardware. OpenAI recently announced work with Broadcom on its own custom inference processor called Jalapeño. Inference refers to the process of using an AI model after it has been trained, such as when a chatbot answers a question.
Amazon and Google also offer their own custom AI chips through their cloud platforms. These developments show that the AI industry is moving towards more specialised hardware, not just bigger models.
What it could mean
If Anthropic and Samsung move forward, the project could help Anthropic secure more control over the technology that powers Claude. It could also add momentum to the wider shift away from total reliance on Nvidia.
For now, the talks remain unconfirmed as a final deal. Still, the discussions show how critical chips have become in the AI sector. As companies compete to build faster, cheaper and more capable systems, custom silicon is becoming a key part of the strategy.