Investors considering Anthropic PBC's planned mega-IPO face an unusual challenge: valuing a company some fear could pose extreme risks to humanity. Researchers on X debated a 10% or greater chance AI could wipe out the human race. Dario Amodei, Anthropic's CEO, outlined concerns about losing control of AI systems, misuse for cyberattacks and bioterrorism, and economic disruption.
While quantifying such risks in an IPO disclosure drew skepticism, the situation mirrors companies handling hazardous materials. A precedent from the 2008 financial crisis emerged, with Slow Ventures partner Sam Lessin proposing the US government force AI labs to place $1 trillion in cash escrow for future cleanup costs. Even this may be insufficient, as Lessin warned liabilities could reach $2 trillion or $4 trillion, making companies worth less than zero.
Some investors may overlook undefined costs based on the theory others will pay the bill. Anthropic declined to comment. The IPO aims to raise as much as or more than Space X's record debut, though the debate turned apocalyptic.
Elon Musk's Space X, whose Grok AI competes with Anthropic, saw shares trade over 20% above its $86.2 billion offering price. President Trump's "morally binding" AI safety accord with tech leaders, including Amodei, calmed some anxiety. Regulators may eventually require safety spending, which could increase costs but enhance customer trust and create a competitive moat against Chinese or open-source competition.
Though insurance coverage limits historically leave shareholders with large disaster bills, as seen with BP's Deepwater Horizon settlement.
Source: Bloomberg Markets · Summarized by HeadlinesBriefing