Anthropic has offered to lease up to $10 billion of computing power over two years from the Meta platform as the AI developer prepares for an October IPO.
Meta is considering the offer after Anthropic presented terms in June, Reuters reported, citing The New York Times. The planned deal would require Anthropic to make monthly payments to access Meta’s computing power.
The report added that both companies could terminate the contract before the two-year period expires. Meta and Anthropic have not finalized the agreement, and its value and duration will depend on the outcome of the talks.
For Anthropic, the lease provides access to the processing power needed to train and operate advanced artificial intelligence models. AI developers rely on large numbers of specialized chips and data centers, making reliable computing access central to their expansion plans.
Meta derives its revenue primarily from the infrastructure built for its AI products and services. The proposed lease would give the social media company another way to generate returns from its computing investments beyond its advertising business, the report said.
Meta may enter the AI infrastructure market
Once the deal closes, Meta will compete with CoreWeave and Nebius, two companies that provide computing infrastructure for AI workloads. Reuters reported that Anthropic’s proposal could turn Meta into a provider of capabilities to outside AI developers while continuing to build its own models and products.
The talks emerge as technology companies compete for chips, power and data center space. Under the reported structure, Anthropic would secure capacity from companies that are spending heavily on AI infrastructure, while Meta would add potential customers for resources within its own computing network.
Anthropic has also made progress on individual long-term infrastructure developments. Earlier this month, the company signed a 20-year data center lease agreement with Bitcoin miner TeraWulf. The deal is expected to provide additional computing resources for Anthropic’s future AI developments.
Taken together, the meta discussion and TeraWulf’s lease show how Anthropic is assembling the necessary infrastructure to support its model. However, evaluation of the size and financial impact of these agreements will depend on their final terms and the amount of capacity that Anthropic ultimately uses.
Anthropic could hit the public market in October
Bloomberg reported that Anthropic is preparing for a possible stock market listing. Banks working on the proposal have begun arranging meetings with company executives and potential investors, the report said.
Those meetings could support an IPO as early as October, but Bloomberg’s schedule remains subject to market conditions and the company’s final decisions. If it debuts in October, Anthropic would go public before OpenAI, but Bloomberg reported that it is considering a 2027 listing.
Chinese AI developer DeepSeek is also preparing for an eventual public offering, according to an original report. Anthropic could therefore be one of the first major companies from the latest generation of AI model developers to go public.
Before the IPO report came out, Anthropic had received approval from the U.S. government last month to restore access to the Mythos 5 model to some companies and federal agencies. Along with the infrastructure deal and investor meetings, the decision adds another development for banks and potential shareholders to consider as preparations progress.
