Meta Platforms is engaged in preliminary discussions with artificial intelligence company Anthropic regarding a substantial infrastructure arrangement that could be valued at up to $10 billion spread across two years, according to reporting from the New York Times citing three unnamed individuals familiar with the negotiations. The arrangement would involve Meta providing crucial computing capacity to Anthropic, marking a strategic pivot for the social media conglomerate as it explores revenue opportunities in the lucrative field of artificial intelligence infrastructure.
The reported talks represent an early-stage exploration by both parties, with neither company confirming the discussions publicly. Meta declined to comment when contacted by Reuters, while Anthropic similarly refrained from making any official statement. Industry observers emphasise that preliminary negotiations of this nature frequently do not progress to final agreements, and the terms under discussion remain fluid and subject to modification. Neither organisation has committed to a specific timeline or binding arrangement at this stage.
Under the proposed structure outlined in the New York Times report, Anthropic would remit monthly payments to Meta throughout the two-year engagement period. Critically, both parties would retain the flexibility to terminate the arrangement ahead of schedule, providing an exit mechanism should circumstances change or either organisation determine that continuing the partnership no longer serves their strategic interests. This flexibility likely reflects the uncertain regulatory environment surrounding artificial intelligence infrastructure deals and the rapidly evolving nature of AI technology adoption.
For Meta, pursuing such a computing lease arrangement aligns with broader strategic objectives articulated by chief executive Mark Zuckerberg during the company's shareholder meeting in May. Zuckerberg indicated at that time that diversifying into cloud computing infrastructure represented a viable path forward, noting that technology firms approached Meta with considerable frequency—nearly every week—seeking access to its surplus computing capabilities or its artificial intelligence models. This commentary suggested that latent demand existed in the marketplace for Meta's infrastructure resources.
The potential Anthropic transaction would position Meta in direct competition with specialised cloud infrastructure firms that have emerged to serve the artificial intelligence sector, including CoreWeave and Nebius. These companies have developed business models centred entirely on providing computing capacity tailored to demanding AI workloads. By monetising its own substantial infrastructure investments, Meta could establish a secondary revenue stream that supplements its core advertising business, which remains vulnerable to regulatory scrutiny and shifting consumer behaviours.
Anthropically initiated the proposal during June discussions with Meta, suggesting the AI company has been actively identifying suppliers capable of meeting its enormous computational requirements. As AI models grow more sophisticated and resource-intensive, organisations like Anthropic face escalating pressure to secure reliable, cost-effective access to computing power. The timing reflects accelerating industry demand for specialised infrastructure to train and operate large language models like Claude, which powers the code-generation capabilities that have gained prominence in developer communities.
Anthropic's approach parallels strategies employed by other technology firms seeking to secure computing capacity. Most notably, the company concluded a separate arrangement with SpaceX in May, gaining access to the entire computing output of SpaceX's Colossus 1 data centre facility located in Memphis, Tennessee. That agreement demonstrated Anthropic's willingness to structure partnerships with infrastructure providers and underscores the sector-wide shortage of available computing resources suitable for advanced artificial intelligence operations. The Colossus facility represents one of the largest dedicated AI computing installations globally.
Meta's consideration of the Anthropic proposal comes amid broader corporate efforts to establish commercial cloud computing businesses. Bloomberg News reported earlier in July that Meta was developing an internal cloud service specifically designed to commercialise excess computing resources and provide hosting capabilities for developers seeking to deploy artificial intelligence models. This initiative suggests Meta views infrastructure monetisation as an important strategic priority moving forward, particularly given the substantial capital expenditures the company has committed to AI research and development.
The negotiations underscore significant structural shifts occurring throughout the technology industry as artificial intelligence adoption accelerates. Large technology companies that developed expansive computing infrastructure primarily to serve internal needs are increasingly exploring opportunities to monetise those assets by becoming computing providers themselves. This transformation reflects both the enormous capital requirements associated with modern AI infrastructure and the potential profitability available to firms that can efficiently deliver computing capacity at scale.
For Malaysian and Southeast Asian technology observers, the reported deal carries implications regarding how regional artificial intelligence development may be constrained by infrastructure scarcity and concentration of computing resources among large American corporations. As developing economies throughout the region seek to build indigenous artificial intelligence capabilities, the dominance of American firms like Meta and Anthropic in controlling critical computing infrastructure highlights a fundamental dependency that may shape technological sovereignty considerations in coming years. The trend toward infrastructure leasing arrangements also raises questions about pricing dynamics and equitable access to computational resources necessary for AI advancement across different markets and economic scales.
