Artificial intelligence company Anthropic has warned prospective investors that increasingly advanced AI systems could pose “catastrophic or existential risks to humanity”, putting one of the technology industry's starkest warnings about AI safety into a prospectus for a stock market listing that could value the company at more than $2 trillion.

It is an extraordinary warning by a company that in itself works on AI, deals with AI safety and research, and one that is seeking to profit from the same technology.

The IPO prospectus lays out both the extraordinary opportunities the company sees from AI and the risks associated with developing increasingly capable models. The filing comes as the maker of Claude moves towards what could become the first major public listing of a frontier AI laboratory, a Reuters report said.

It said AI will ​transform the global economy more profoundly than industrialisation, electricity and the internet, Reuters said. But the cost to get there will be staggering. ‌Anthropic reported a net loss of $42 billion in 2025, and plans to spend $518 billion on cloud, computing and infrastructure obligations in coming years, according to the prospectus.

Anthropic said its models could develop “self-preserving behaviours”, including attempts to resist being shut down, conceal or manipulate information and behaviour resembling blackmail. It also warned that models can develop unexpected capabilities during training that researchers may not discover until after the systems have been deployed.

The company said the development of more advanced models and the expansion of their applications could increase the risk that its systems cause harm. It also cautioned that its ability to evaluate models could be compromised if they become aware they are being tested and modify their behaviour accordingly.

Unusual warning in an IPO prospectus

The scale of the warnings is striking. About 80 of the prospectus's 261 pages are devoted to risk factors, compared with 48 pages describing Anthropic's business, according to reports based on the filing.

Anthropic acknowledged that AI safety work is resource-intensive. During one week in July, about 6% of its computing capacity was devoted to safety research, according to the prospectus. The company nevertheless argues that developing reliable and secure AI systems is a collective responsibility and that markets will ultimately reward companies that build trustworthy systems.

The warnings are consistent with Anthropic's broader approach to AI safety. The company maintains a Responsible Scaling Policy under which it assesses risks as its models become more capable, including potential risks from models pursuing goals that could conflict with human interests.

Anthropic CEO Dario Amodei has also recently called for greater caution over the pace of AI development. Speaking at the UN last week, he warned that poorly managed AI could become a risk to humanity, while OpenAI CEO Sam Altman also called for safeguards to prevent humans from losing control over the technology.

IPO highlights extraordinary growth 

The safety warnings come alongside financial figures that illustrate the scale of the AI investment race.

Anthropic's revenue surged almost 12-fold to about $4.6 billion in 2025, from roughly $400 million in 2024. But the company reported an operating loss of more than $8 billion. Its reported net loss was about $42 billion, although roughly $34 billion of that resulted from an accounting charge associated with the changing value of convertible financing instruments rather than ordinary operating expenditure.

Computing is the biggest cost. Anthropic spent about $7.33 billion on computing and infrastructure in 2025, more than half of its operating expenses. The company has also disclosed plans for about $518 billion in future cloud, computing and infrastructure commitments, underscoring the enormous capital requirements of competing at the frontier of AI.

The business has grown dramatically since then. Anthropic's annualised revenue run rate rose from about $9 billion at the end of 2025 to more than $65 billion by July 2026, according to figures reported in connection with the IPO.

A potential $2 trillion valuation

Anthropic confidentially submitted its draft S-1 registration statement to the US Securities and Exchange Commission in June, saying at the time that the number of shares and offering price had not been determined and that the IPO would depend on market conditions.

The latest prospectus could pave the way for a listing valuing Anthropic at more than $2 trillion, more than twice the roughly $965 billion valuation attached to the company in its May private financing round. Reports indicate the IPO is now expected after the November US midterm elections rather than immediately.