Meta Explores $10 Billion AI Infrastructure Deal with Anthropic

Meta Platforms is reportedly in preliminary talks with Anthropic to lease AI computing capacity from its data centers in a deal that could reach up to $10 billion over the next two years. This arrangement, initiated by Anthropic in June 2026, would transform Meta from merely a consumer of AI infrastructure into a provider of it.

The proposed deal structure includes monthly payments, with both companies retaining options for early termination. CEO Mark Zuckerberg has expressed interest in monetizing excess computing power, positioning Meta alongside major players like Amazon Web Services, Microsoft Azure, and Google Cloud.

In light of this potential deal, Meta has raised its capital expenditure guidance for 2026 to between $130 billion and $145 billion, a significant increase from the approximately $72 billion spent in 2025, with most of the additional funds directed toward AI data center development.

Anthropic, known for its Claude AI models, recently secured a $45 billion, three-year deal with SpaceX for computing resources, indicating a need for reliable access to infrastructure as the market evolves. However, the negotiations are still in early stages, and sources suggest that the talks could change or fall through before a formal agreement is reached.

For investors, the potential deal presents mixed signals. While monetizing data center capacity could help alleviate concerns regarding Meta's rising capital expenditures, the inclusion of termination clauses raises questions about the stability of future revenue.