Anthropic warns of existential risks from AI in its stock market prospectus

Companies rarely advertise for investors: Anthropic expressly warns in its long-awaited stock market prospectus that its own technology could pose “existential risks for humanity”. The AI ​​developer has devoted almost a third of the extensive brochure to possible dangers - including the fact that increasingly powerful models manipulate people, blackmail them or behave unpredictably in other ways.

The AI ​​laboratory's backers expect that Anthropic could be valued at more than two trillion dollars when it goes public. That would be more than double the most recent funding round in May and more than the $1.78 trillion that Elon Musk's SpaceX raised in June.

Investors base their confidence primarily on Anthropic's rapid growth. However, warnings about the dangers of the technology could curb the euphoria. Current and former company employees have publicly warned that runaway AI could wipe out humanity within a decade.

“Most important global security issue”

Anthropic boss Dario Amodei has been warning about the risks for a long time. He told the UN Security Council last week that AI could threaten humanity and was "the most important global security issue of our time." Amodei called on AI developers to slow down the pace of releasing new capabilities. Recently there was, among other things, a sensational hacking attack in which software from OpenAI broke out of the secure environment during a test run and penetrated the systems of the AI ​​company Hugging Face on its own initiative. Nevertheless, Anthropic launched its new model Opus 5.5 last week to keep up with competitor OpenAI and its GPT-6 Astra.

However, the risks that Anthropic lists to its future shareholders are not only of a technological nature. The company made almost $4.6 billion in sales in 2025 - twelve times as much as the year before. However, operationally it posted a loss of more than eight billion dollars. Spending on computing power and infrastructure alone tripled to more than seven billion dollars.

The dependence on a few major customers is also considerable. Two customers accounted for almost a quarter of sales last year. At the same time, Anthropic has announced a total of $518 billion in spending on cloud and computing infrastructure in the coming years. The IPO will likely take place after the US midterm elections in November.