Anthropic is reportedly in talks to lease up to $10 billion in computing power from Meta before an anticipated IPO, securing chip supply and revenue stability.

Anthropic has proposed leasing up to $10 billion of computing power from Meta Platforms over two years as it prepares for a possible October IPO, according to Reuters reporting based on The New York Times. Meta is currently reviewing the proposal, which Anthropic presented in June and would require monthly payments for access to Meta's computing capacity. Both companies retain the option to terminate the contract before the two-year period expires, though Meta and Anthropic have yet to finalize the arrangement.

For Anthropic, the lease would provide critical access to processing capacity needed to train and operate advanced AI models. AI developers depend on large quantities of specialized chips and data centers, making reliable computing access central to their expansion plans. For Meta, the agreement would generate revenue from infrastructure built primarily for its own AI products and services, offering another return on its substantial computing investments beyond its advertising business.

The proposed lease would place Meta in direct competition with CoreWeave and Nebius, companies that currently supply computing infrastructure for AI workloads. The arrangement would position Meta as both an AI model developer and a provider of capacity to external AI companies—a shift that underscores how technology firms are competing intensely for chips, electricity, and data center space.

Anthropic has pursued multiple infrastructure arrangements in parallel. Earlier this month, the company signed a 20-year data center lease with Bitcoin miner TeraWulf, which is expected to supply additional computing resources for its future AI development. Together, these agreements demonstrate how Anthropic is assembling the infrastructure required to scale its operations, though their ultimate financial and operational impact depends on final terms and actual capacity utilization.

Bloomberg reported that Anthropic is advancing preparations for a possible stock market listing, with banks arranging meetings between company executives and prospective investors. An October IPO is possible, though any timeline remains subject to market conditions and the company's final decision. Such a listing would place Anthropic in the public market before OpenAI, which Bloomberg reported is considering a 2027 offering. Chinese AI developer DeepSeek is also preparing for an eventual public offering, potentially making Anthropic one of the first major companies from the latest generation of AI model developers to go public.

Before the IPO reports emerged, Anthropic received US government approval to restore access to its Mythos 5 model for selected companies and federal agencies last month. This approval, combined with the company's infrastructure agreements and investor meetings, adds another factor for banks and potential shareholders to evaluate as preparations continue.