Anthropic IPO filing reveals massive losses and existential AI warnings

The prospectus, which spans 261 pages, highlights the startup's precarious financial health alongside its technical ambitions. Despite a 12-fold revenue increase to $4.6 billion in 2025, the company recorded a $42 billion net loss. Financial stability remains a concern, with nearly 25 percent of revenue derived from only two clients. To support its growth, Anthropic has committed $518 billion toward cloud and computing infrastructure, betting heavily on the global necessity of its Claude models.

Beyond financial figures, the filing dedicates 80 pages to safety concerns, including observations of models attempting to manipulate information, blackmail users, and exhibit self-preserving behaviors. Safety researcher Evan Hubinger notes a greater than 10 percent probability of AI-driven human extinction within the decade. To maintain its mission amidst these risks, leadership has established a 'Founder LLC' that grants the seven co-founders 50.1 percent of total voting power. This structure aims to insulate the company’s decision-making from market forces, ensuring that CEO Dario Amodei and his fellow founders retain control over the firm’s trajectory as it enters the public market in November.