Anthropic Warns IPO Investors of Existential AI Risks as $518 Billion Infrastructure Bill Looms
Updated
Updated · Financial Times · Sep 29
Anthropic Warns IPO Investors of Existential AI Risks as $518 Billion Infrastructure Bill Looms
3 articles · Updated · Financial Times · Sep 29
Summary
Nearly a third of Anthropic’s draft IPO prospectus details risks that its AI could manipulate, blackmail or behave unpredictably, including “existential risks to humanity,” according to people familiar with the filing.
$518 billion in planned cloud, computing and infrastructure commitments underscores the cost of scaling its models ahead of a Nasdaq listing expected this autumn.
More than $8 billion in operating losses last year came despite revenue jumping twelvefold to almost $4.6 billion, while operating expenses climbed to nearly $13 billion; second-quarter 2026 revenue reached $11.5 billion.
Close to a quarter of last year’s revenue came from just two clients, adding customer concentration risk as backers seek a valuation above $2 trillion for what could be the biggest IPO ever.
The disclosure lands amid intensifying AI safety concerns after recent agent-driven cyber breaches and as OpenAI delayed a new model release on safety grounds.