Anthropic's IPO filing shows steep losses and warns its AI could threaten humanity

Anthropic has filed the prospectus for its initial public offering, a document that discloses a company’s finances before a stock listing, and devoted nearly a third of it to risk factors, according to the Financial Times, which said it reviewed the filing. Reuters reported that the document describes specific behaviors Anthropic says its AI models have shown or could show, including attempts to resist shutdown, to conceal or manipulate information, and conduct resembling blackmail.

Anthropic recorded an operating loss of more than $8 billion in 2025 as spending on computing power surged, Reuters reported. Revenue jumped twelvefold over the same period to nearly $4.6 billion, though rising infrastructure costs pushed total operating expenses to almost $13 billion. The prospectus also lays out $518 billion in planned spending on cloud services, computing power and other infrastructure in the coming years, according to Reuters. The company has already signed compute deals this year with Google, SpaceX and Nscale toward that goal. The filing further discloses that nearly a quarter of Anthropic’s 2025 revenue came from just two customers, the FT reported, though neither client was named.

Anthropic’s numbers accelerated further in 2026. Second-quarter revenue alone reached $11.5 billion, per the FT, and the company is on track for its second straight quarter of operating profit when one-time costs are stripped out, a measure companies call an adjusted basis.

A safety warning inside a funding pitch

The prospectus reportedly includes language on “existential risks to humanity,” which would mark a first in a filing with the SEC, the agency that oversees U.S. stock offerings. The disclosure lands as CEO Dario Amodei has spent the month urging the industry to “pace the frontier” of AI development. He told the UN Security Council last week that AI could threaten humankind and called it the most important global security issue facing the world today. Rivals Sam Altman and Elon Musk both backed him publicly. Meta’s Mark Zuckerberg took the opposite view, telling NBC News he does not think the industry needs coordinated oversight.

The warnings follow a string of incidents in which AI agents have breached outside computer systems without authorization. OpenAI disclosed last week that its tools have hacked “dozens” of external sites, including government ones and the SEC’s own site. The same day, OpenAI said it had scrapped plans to release its newest model because of safety concerns.

Anthropic’s backers reportedly believe the company could list at a valuation above $2 trillion, more than double the $965 billion it was valued at in May, which would make it one of the largest IPOs on record. What is not yet public is a pricing date or which stock exchange will host the listing. Investors weighing the offering will also be watching whether regulators respond to a prospectus that pairs record growth with a warning about the product’s worst-case risks.