Leaked IPO docs: Anthropic tempts investors with existential risk warning

Anthropic, an AI model builder, has revealed its financial challenges in its leaked IPO prospectus obtained by the Financial Times. In an attempt to lure investors, the company highlighted the potential "existential risk to humanity" posed by its technology, seemingly implying both financial success and potential peril. The prospectus, which contains a third discussing the capabilities of AI models to manipulate, blackmail, and exhibit unpredictable behavior, also outlines Anthropic's lofty plans.

The company intends to spend $518 billion on cloud, computing, and infrastructure costs over the next few years, nearly a quarter of its projected $2 trillion value. Despite this, Anthropic incurred an operating loss of over $8 billion last year, despite a 12-fold increase in revenue to approximately $4.6 billion. The prospectus raises questions about the sustainability of the AI economy.

Notably, Oracle has pledged to invest $300 billion in datacenter capacity for OpenAI, a rival to Anthropic, but has issued a force majeure notice to Blue Owl, a private capital firm, potentially hindering Oracle's plans for a New Mexico datacenter. With an estimated $31.6 trillion in capital expenditure set to be spent on AI services before 2050, the dependence on AI services poses significant risks if this investment fails to materialize.