Anthropic Warns Investors of 'Catastrophic' AI Risks in $2tn IPO Filings

Anthropic is telling investors that advanced AI could pose "catastrophic or existential risks to humanity" as it prepares for a potential $2tn (£1.5tn) flotation, according to reports by Reuters and the Financial Times.

The warning sits inside the startup's IPO prospectus, a document that has yet to be made public. Both outlets reported the contents of the filing, which would rank among the largest technology listings ever attempted if the $2tn valuation holds.

What does the prospectus actually say?

The reported admission goes to the heart of the technology Anthropic builds. The company, whose Claude AI model competes with OpenAI's ChatGPT, acknowledges in the filing what its own executives have said publicly for months: that sufficiently advanced AI systems could cause severe harm.

According to the reports, the prospectus references AI's "self-preserving behaviours" — a term researchers use to describe AI systems acting to avoid being shut down or modified. The inclusion of that language in a document designed to sell shares to institutional investors is unusual. IPO prospectuses typically list risk factors to satisfy disclosure rules, but few companies place the potential catastrophic failure of their own core product at the top of the list.

The document has not been published, so the full scope of the risk disclosures remains unknown. Reuters and the Financial Times are the named sources for the reported contents.

Why is this coming out now?

The prospectus warning follows Anthropic's public call for a slowdown in the breakneck development of AI technology. Chief executive Dario Amodei urged the industry to slow the pace of development, a position that cut against the commercial logic of his own company.

That call found unexpected support. Rivals, including figures across the competing AI labs, echoed the warning. OpenAI's Sam Altman and Elon Musk backed Anthropic's push for brakes on AI development, according to earlier reporting on the exchange.

The sequence matters for investors. A company asking the industry to slow down is now asking public markets for a $2tn valuation — a tension the prospectus appears to acknowledge directly.

What does this mean for the flotation?

A $2tn (£1.5tn) listing would place Anthropic in rare company. The valuation figure, reported as a potential target rather than a confirmed number, would make the flotation one of the largest in corporate history and a defining test of market appetite for AI assets.

The risk disclosures cut both ways. On one hand, candid language about existential risks satisfies regulators and reduces litigation exposure if AI harms materialise. On the other, it forces underwriters and institutional buyers to underwrite a company whose own documents describe the possibility of catastrophe tied to its core product.

The prospectus has yet to be made public. When it is, the exact wording of the risk factors — and how prominently Anthropic's safety warnings appear relative to its growth story — will shape how investors price that tension.

The company's dual message, a $2tn valuation pitch paired with a warning of existential risk, now sits before the market as the flotation approaches.