The proposed two-year agreement would allow Anthropic to lease Meta’s computing infrastructure through monthly payments while giving the social media company a new source of revenue beyond advertising.
By yourNEWS Media Newsroom
Meta Platforms is discussing a potential agreement to provide Anthropic with as much as $10 billion in computing capacity over two years, according to a report citing three people familiar with the negotiations.
Anthropic proposed the arrangement in June, and Meta is considering the offer, The New York Times reported, according to the people briefed on the discussions.
Under the proposed structure, Anthropic would make monthly payments to Meta for access to its computing infrastructure. The terms could still change, and both companies would reportedly have the option to end the agreement before the two-year period expires.
Reuters said it could not independently verify the report.
Anthropic declined to comment, while Meta did not immediately respond to a request for comment.
Meta shares slightly reduced their losses following publication of the report but remained down nearly 3% amid a broader technology-sector selloff.
A completed agreement would give Meta an opportunity to generate additional revenue from the extensive computing infrastructure it has built to develop and operate artificial intelligence systems.
Meta currently relies heavily on digital advertising for revenue. Leasing computing capacity to outside artificial intelligence companies would help diversify its business while allowing the company to recover more value from investments in data centers, chips and related infrastructure.
The deal would also place Meta in more direct competition with companies including CoreWeave and Nebius, which specialize in supplying the high-performance computing capacity required to train and run advanced artificial intelligence models.
Demand for that infrastructure has risen as businesses and consumers increasingly use sophisticated AI systems that require large amounts of processing power.
Anthropic develops the Claude family of artificial intelligence models and Claude Code, a tool designed to assist with software development.
The company would use Meta’s infrastructure to support the growing computing demands associated with training models, serving customers and operating its expanding range of AI products.
The potential agreement illustrates how the artificial intelligence industry is creating new relationships among companies that may compete in one area while doing business together in another.
Meta develops its own AI models and consumer products but may still benefit financially by leasing unused or available infrastructure to Anthropic, another major model developer.
For Anthropic, the arrangement could provide access to substantial computing capacity without requiring the company to build all of the necessary infrastructure itself.
The negotiations remain preliminary, and neither company has confirmed that a final agreement will be reached.
Original article
Comments