Anthropic's IPO Filing: $4.6B Revenue, $8B Loss, $2T Ambition

Anthropic warns in its own IPO prospectus that the AI technology it builds could pose "existential risks to humanity." The company put that language in the document it is using to persuade public investors to buy its shares — a striking admission for a firm whose backers are seeking a valuation above $2 trillion.

The numbers in the filing tell a story of extraordinary growth and extraordinary cost. Revenue grew twelvefold in 2025 to $4.6 billion, according to the prospectus. In the same period, the operating loss widened to $8.06 billion. Anthropic is scaling its business faster than almost any software company in history, and it is still spending nearly twice its annual revenue to do so.

What the filing shows

Three figures anchor the filing.

First, revenue: $4.6 billion for 2025, up twelvefold from the prior year. That growth rate places Anthropic among the fastest-scaling companies in the modern software industry, and it reflects surging demand for its Claude models across consumer subscriptions and enterprise contracts.

Second, the loss: $8.06 billion in operating losses for the year, and widening. The gap between what Anthropic earns and what it spends reflects the core economics of frontier AI development. Training and running large models consumes billions of dollars in computing power, and the prospectus numbers show that revenue growth alone has not yet closed that gap.

Third, the valuation target: backers are aiming for a valuation above $2 trillion, according to the filing. If achieved, that would place Anthropic among the most valuable companies in the world — before it has turned an operating profit.

The "existential" language

The most unusual element of the filing is its risk language. Anthropic states in the prospectus that its AI technology could pose "existential risks to humanity."

The phrasing is notable on two levels. It puts the company's own stated safety concerns into a formal legal document for public investors. And it asks those investors to buy shares in a company that explicitly flags the possibility that its products could, in its own words, threaten humanity.

Companies routinely list risk factors in IPO documents — competition, regulation, key-person departures. A frontier AI lab naming the existential stakes of its own technology as a risk factor is a different order of disclosure. It reflects the dual position these companies occupy: they argue their models are transformative enough to justify trillion-dollar valuations, while acknowledging in the same document that the same capabilities carry risks they say they take seriously.

For Anthropic specifically, the language is consistent with the company's public positioning since its founding. It has long framed safety research as central to its mission, and the prospectus now carries that framing into the capital markets.

Why the stakes are high for the industry

The IPO is expected to set the bar for the entire AI industry, according to the filing's context. That makes this more than a single company's listing.

Anthropic's public debut will serve as the first direct test of how public markets value frontier AI development at scale. The offering will establish a benchmark that other AI labs, their investors, and their bankers will reference when pricing subsequent listings and private rounds. A valuation above $2 trillion would signal that public investors accept the current economics of frontier AI — massive compute spending funded by massive capital raises, with profitability deferred indefinitely.

The loss figure matters as much as the revenue figure in that calculation. An $8.06 billion operating loss against $4.6 billion in revenue means Anthropic's costs run at roughly 1.75 times its income. Investors buying into the IPO are underwriting the bet that revenue growth continues to outpace the growth in compute costs, and that at some point the two lines cross. The twelvefold revenue increase in 2025 is the strongest evidence for that bet. The widening loss is the strongest evidence against it.

The valuation target compounds the stakes. At above $2 trillion, Anthropic would be priced at more than 400 times its 2025 revenue — a multiple that leaves little room for any slowdown in growth.

What to watch

The filing itself moves the story forward. The specific terms — share price, offering size, listing venue — will follow as the IPO process advances toward a public debut.

The numbers investors will scrutinize are already on the table: $4.6 billion in revenue, an $8.06 billion operating loss, twelvefold growth, and a $2 trillion-plus ambition. How public markets price that combination will shape the financing environment for every frontier AI lab that follows Anthropic to market — and will determine whether the industry's current burn rate can sustain itself on public capital rather than private backing.