Meta Platforms is reportedly in discussions with artificial intelligence startup Anthropic over a computing infrastructure agreement that could be worth as much as $10 billion over two years, marking what could become one of the largest AI compute leasing deals to date. If finalised, the agreement would represent a strategic shift for Meta as it seeks to monetise its growing AI infrastructure while supporting Anthropic's increasing demand for computing power.
According to people familiar with the discussions, Anthropic proposed the arrangement in June, with Meta currently evaluating the terms. The proposed structure would see Anthropic make monthly payments over a two-year period for access to Meta's computing infrastructure. The discussions remain preliminary, and both companies would have the option to terminate the agreement early if a final contract is reached. Neither Meta nor Anthropic has publicly confirmed the negotiations.
The potential partnership highlights the growing importance of computing infrastructure in the AI industry. As companies develop increasingly sophisticated foundation models, demand for graphics processing units (GPUs), data centres and high-performance computing clusters has intensified, making access to compute one of the industry's most valuable resources. AI developers are now investing heavily in securing long-term computing capacity to train and deploy advanced models.
For Meta, the agreement could open a new revenue stream beyond its core advertising business. The company has committed between $125 billion and $145 billion in capital expenditure this year, with a significant portion allocated to AI infrastructure, including next-generation data centres and specialised computing clusters. Leasing excess capacity could allow Meta to generate returns from these investments while expanding into AI cloud services, an area currently dominated by providers such as Amazon Web Services, Microsoft Azure and Google Cloud.
The reported discussions also reflect a broader industry trend in which frontier AI developers increasingly rely on multiple infrastructure partners rather than a single cloud provider. Anthropic already works with several major technology companies for computing capacity and cloud services, as competition for AI infrastructure continues to intensify. Diversifying compute providers could help AI companies improve resilience while ensuring access to the resources needed to train increasingly complex models.
Meta has previously indicated that it could consider offering computing services to external organisations if demand justified such a move. Chief Executive Officer Mark Zuckerberg has spoken about the company's expanding infrastructure investments and suggested that surplus computing capacity could eventually be made available to third parties. Reports also indicate that Meta has established a dedicated "Meta Compute" initiative as part of its broader AI infrastructure strategy.
The emergence of AI compute as a commercial business reflects the rapid evolution of the artificial intelligence ecosystem, where access to infrastructure is becoming as strategically important as advances in models themselves. As enterprises, developers and AI startups compete for limited computing resources, infrastructure providers are increasingly exploring new business models centred on leasing high-performance compute.
While the proposed Meta-Anthropic agreement remains under discussion, it underscores how AI infrastructure is becoming a critical battleground in the next phase of artificial intelligence development. Should the deal proceed, it would further reinforce the growing commercial value of AI compute as technology companies race to meet surging global demand for advanced computing capacity.