Meta is reportedly in early discussions to lease its computing power to AI company Anthropic in a potential deal worth up to $10bn over two years. Anthropic, known for its Claude chatbot, is seeking additional computing resources to meet growing demand and serve its AI models to clients.

This prospective agreement would involve monthly payments over the two-year period. The talks are in their initial stages, and a definitive agreement is not assured. The deal comes as Meta plans to launch its own cloud-computing business to monetize excess capacity, a move that would require significant upfront investment and competition with established players like Amazon, Microsoft, and Google.

Meta has invested heavily in its data centre infrastructure, with $145bn earmarked for capital expenditures this year, primarily for AI. This strategy includes developing its own AI models to rival those from Anthropic, OpenAI, and Google. The potential deal with Anthropic highlights the immense demand and high costs associated with developing and deploying advanced AI technologies, with companies like SpaceX also making substantial payments for computing power.