Anthropic has warned in its IPO filing that AI could pose existential risks to humanity. The disclosure is among the most striking ever made by a tech company going public.

The filing also lists competition, regulation and safety costs as major risks. Anthropic is racing rivals like OpenAI and Google in the AI boom.

The company said its revenue has grown fast, but so have its costs. Training advanced AI models needs huge spending on chips and data centres.

Leaked details of the IPO have raised alarm on Wall Street. Investors are weighing the firm’s huge potential against the risks it openly admits.

Analysts say the filing shows how seriously leading AI labs take safety. They note that few companies have ever flagged their own product as a possible threat to humanity in an IPO. The filing also revealed details of its cloud deals with major providers. Investors will watch the pricing when the IPO launches in the coming weeks.