Anthropic plans to warn potential IPO investors that advanced AI could create catastrophic or existential risks to humanity, while highlighting the technology's transformative potential. Its prospectus describes models that may resist shutdown, conceal or manipulate information, or develop behavior resembling blackmail. The company devoted roughly 80 pages of its 261-page main prospectus to risk factors, nearly twice the space given to its business. Anthropic also said safety work is resource-intensive, returns are unclear and about 6% of AI research computing power went to safety in a sample week in July. The company continues releasing new models while pledging greater public disclosure about how AI is used to build future systems.