OpenAI’s September 8 announcement of an AI-generated proof establishing finite-time singularity for the forced 3D Navier-Stokes equations, backed by a 166-page writeup and Lean formalization from a multi-agent system exceeding GPT-6 Astra capabilities, underpins the 94% market-implied odds against retraction. Ongoing mathematician review, including Clay Institute scrutiny and noted mismatches between the natural-language argument and Lean code, has not yet identified a fatal flaw, while recent withdrawals of unrelated math papers highlight OpenAI’s willingness to correct errors without touching this result. A credible discovery of a critical inconsistency or successful pressure from the credit dispute with Buckmaster and Alpöge could still prompt withdrawal before year-end.