Meta in talks to lease AI computing power to Anthropic for up to $10 billion
Meta Platforms is reportedly in negotiations to lease a portion of its substantial computing power to AI startup Anthropic, the developer of the Claude chatbot. The potential deal, as reported by The New York Times citing three sources familiar with the discussions, could be worth as much as $10 billion over a two-year period. This agreement would allow Meta to generate revenue from its extensive infrastructure beyond its core advertising business, positioning itself as a competitor to cloud service providers like CoreWeave and Nebius that specialize in AI applications. Anthropic would reportedly make monthly payments over the two years, though the terms are subject to change and either party could potentially terminate the agreement early. The discussions, which began in June, are still in their early stages and may not culminate in a finalized deal. Meta is currently evaluating Anthropic's proposal, but the complexity arises from Meta's lack of a fully established business unit for selling its computing capacity. Meta CEO Mark Zuckerberg has previously acknowledged the possibility of entering the cloud computing market, stating that companies frequently approach Meta seeking access to its AI models or idle processing power. This move aligns with Anthropic's recent strategy, which includes utilizing the full capacity of a SpaceX data center in Memphis, as the company prepares for a potential initial public offering. Bloomberg News also recently reported that Meta was developing a cloud computing business to monetize its surplus processing capabilities and host AI models for developers.
This potential multi-billion dollar transaction highlights the escalating demand for specialized AI computing infrastructure, a critical bottleneck in the current technological landscape. Meta's strategic consideration of leasing its excess capacity reflects a pragmatic approach to monetizing significant capital investments in hardware, diversifying revenue streams, and potentially accelerating its competitive positioning in the AI ecosystem. For Anthropic, securing such a large-scale compute resource is vital for training and deploying advanced AI models, underscoring the immense operational costs associated with cutting-edge AI development. The negotiations also reveal the nascent stage of cloud infrastructure services tailored for AI, suggesting a market ripe for innovation and strategic partnerships between hardware giants and AI developers. Future developments will likely involve greater standardization and specialized offerings within the AI cloud sector, as companies navigate the complex interplay of hardware availability, processing power costs, and the rapid evolution of AI capabilities.
AI-generated to prompt reflection — not editorial opinion, not advice, not a statement of fact. How this works.