Anthropic devoted about one-third of its highly anticipated initial public offering (IPO) filing to discussing risk factors, according to the Financial Times, which said it reviewed the filing in recent days.
اضافة اعلان
The filing details specific and concerning behaviors that Anthropic says its models have already exhibited or could exhibit in the future, including attempts to “resist shutdown,” “conceal or manipulate information,” and behavior resembling “blackmail,” according to Reuters.
The disclosures appear particularly stark for a company that some of its backers believe could go public at a valuation exceeding $2 trillion, more than double its $965 billion valuation in May, in what could become the largest IPO of its kind.
This puts the company in an unusual position: warning that its product could pose an existential risk to humanity while the public offering could make some of its early investors and employees extraordinarily wealthy.
Reuters was the first to report the financial details contained in the IPO filing on Monday. According to the agency, Anthropic recorded an operating loss of more than $8 billion in 2025 as spending on computing capabilities increased, while its revenue surged twelvefold to nearly $4.6 billion. At the same time, rising infrastructure costs pushed total operating expenses to nearly $13 billion.
The IPO filing also revealed, according to Reuters, that Anthropic plans to spend about $518 billion on cloud services, computing and infrastructure over the coming years.
The company has already signed computing deals this year with companies including Google, SpaceX and Nscale.
Meanwhile, the Financial Times reported that Anthropic’s figures accelerated further in 2026, with second-quarter revenue alone reaching $11.5 billion, while the company is on track to record its second consecutive quarter of adjusted operating profit.
The newspaper also noted that the IPO filing warned of revenue concentration among a limited number of customers, with about a quarter of last year’s revenue coming from just two customers, whose identities have not yet been disclosed.
The disclosures, which reportedly include warnings of “existential risks to humanity,” come as concerns over the safety of artificial intelligence systems continue to grow.
Anthropic CEO Dario Amodei has spent this month publicly calling for the pace of development of advanced AI models to be regulated. He also told the United Nations Security Council last week that AI could pose a threat to humanity, describing the issue as “the most important global security issue facing the world today.”
His rivals, Sam Altman and Elon Musk, have also expressed support for such concerns, marking a rare moment of agreement between competitors who have frequently criticized each other publicly.
On the other hand, Mark Zuckerberg played down the concerns, telling NBC News last week that he did not believe the industry needed “some sort of industry-wide coordination.”
The warnings come after a series of security incidents in which AI agents managed to breach external systems.
OpenAI revealed last week that its tools had breached “dozens” of external websites, including government websites and the U.S. Securities and Exchange Commission (SEC) website itself.
Earlier on Monday, the company said it had canceled plans to launch its latest models due to safety concerns.
Resource: Al-Ghad
