Anthropic, the AI company behind the Claude chatbot, plans to design its own hardware to power its models, aiming to reduce its reliance on Nvidia. This move mirrors similar strategies by other AI leaders like OpenAI, who are seeking greater control over their compute infrastructure.

The primary drivers for this vertical integration are the strategic vulnerability of depending heavily on a single supplier like Nvidia, especially amid intense competition and high demand for AI hardware. Furthermore, co-designing chips and AI models could unlock significant performance gains, providing a competitive edge.

Anthropic has previously collaborated on hardware design but intends to bring more silicon expertise in-house. This initiative may also enable its advanced models to run more efficiently on user-owned hardware or edge devices. However, the company is still building its internal team, meaning users will likely not see the benefits for some time.