Anthropic is moving ahead with a blockbuster I.P.O. even as C.E.O. Dario Amodei calls for slowing advanced A.I. development. The company is expected to generate over $100 billion in annualized revenue this year, up from $65 billion in July, justifying a potential $2 trillion valuation.

Financial documents could be released within weeks, with trading possibly starting in November. The offering highlights a tension: despite safety concerns, A.I. returns are tantalizing. Going public would fund substantial computing costs.

Investors expect Anthropic to have about five gigawatts of compute by year-end, doubling next year, rivaling OpenAI. On Thursday, Anthropic hosted about 100 investors at its San Francisco headquarters. Co-founders Jared Kaplan and Ben Mann, Claude Code head Boris Cherny, and researcher Andrej Karpathy presented.

Dr. Amodei, C.F.O. Krishna Rao, and bankers have met with potential investors. Some smaller investors haven't secured meetings. Anthropic has pitched A.I.'s positive aspects, including biology research and curing diseases, while downplaying threats from cheaper open-weight models.