Commentary: Anthropic’s record US$2 trillion stock market listing highlights growing risks for us all

Anthropic, the developer of AI chatbot Claude, is set to launch one of the largest initial public offerings (IPOs) of all time, valued at up to US$2 trillion. This would surpass SpaceX's previous record of US$1.77 trillion. The AI-powered IPOs reflect an increasingly significant trend, with AI risks becoming unavoidable for most investors.

Even a broad index fund like the S&P 500, which holds AI companies, accounts for about 44% of its gains. In Australia, nearly A$1 trillion (US$702.3 billion) was invested in US shares by the end of 2025, including superannuation and pension funds. The bond market is also heavily influenced by AI companies, which have issued over US$466 billion in bonds this year, four times the previous year's pace.

While these companies may appear safe with investment-grade ratings, a simultaneous default in case of an AI market crash could lead to a significant loss. Furthermore, the risk of AI agents misbehaving and causing catastrophic or existential risks to humanity is a growing concern, with companies themselves warning about this in their risk disclosures.

The AI 2027 scenario suggests a feedback loop where AI models improve at a faster rate than humans can check them, potentially leading to a dangerous situation. Experts argue that clear rules and legal obligations are necessary to manage these risks, both at the societal level and within financial markets.