Anthropic’s investors are sketching out a $2 trillion public debut. The market may instead ask whether the Claude maker’s explosive growth can survive government pressure and a rapidly tightening fight over AI prices.

The bullish case gathered force after Anthropic’s annualized revenue passed $47 billion in May and fresh investment pushed its valuation to $965 billion. Backers now expect an autumn flotation and project $100 billion to $120 billion in annualized revenue by the end of 2026 — a scale that could make it the largest IPO ever.

Their arithmetic is unapologetically aggressive. “If Anthropic is growing 800 percent a year, you’d think at the incredibly low end they would trade at 30 times [revenue],” one investor said. “That would make them a $3 trillion company.”¹

But the momentum has already shown its vulnerability. In June, Commerce Department export controls briefly forced Anthropic to withdraw its leading Fable 5 and Mythos 5 models. Investors with knowledge of the company said the disruption contributed to slower revenue growth that month, while some customers were unsettled by the episode.

The political conflict runs deeper than a temporary restriction. Anthropic remains in litigation with the Defense Department after the agency labeled it a supply-chain risk, even as the company tries to widen its lead over OpenAI and Google in business AI sales.

Then comes the commercial squeeze. Anthropic’s top model costs more than two-and-a-half times OpenAI’s flagship, while improving Chinese open-weight models sell for a fraction of the price. Ramp found Anthropic gained US business market share last month, but also concluded that companies were “hitting their limit on AI spend” and moving toward cheaper options.¹

For believers, that is a manageable dent rather than a broken thesis. “It’s easy to come up with challenges,” said another investor, arguing Anthropic remains ahead in performance, positioning and customer appeal.¹ For public-market investors, however, the IPO would turn that confidence into a far more exacting quarterly test.