AI Giant's $800Bn Compute Bet Sparks IPO Frenzy

Anthropic is poised for a significant initial public offering, with its prospectus revealing over 300 pages of its 'mission' to develop AI for humanity's benefit. This venture, however, comes with warnings of "catastrophic or existential risks." The AI lab has secured commitments exceeding $800bn for data centre spending, extending through 2045, highlighting computing power as a critical input.

Financially, Anthropic's revenue has surged, but it relies on a small number of major clients. The company's governance structure is described as "novel," significantly limiting ordinary shareholders' power while granting founders substantial control. This unusual setup prioritizes "the public good" alongside profit, potentially leading to decisions that diverge from investor expectations.

The S-1 filing also dedicates considerable attention to a wide array of risk factors, including the potential for AI to pose existential threats or aid malicious actors. Anthropic acknowledges the inherent unpredictability of its technology, including potential self-preserving behaviors in AI models and the challenge of insuring against such profound risks.