The recent OpenAI announcement of an AI-generated proof for a Navier-Stokes singularity—produced by thousands of autonomous agents and Lean-verified—has lifted trader confidence that large language models can tackle Millennium Prize Problems, yet the 30.5% implied probability tie between Hodge Conjecture and no full solution by December 31, 2027 highlights persistent uncertainty over verification timelines and prize eligibility. Partial AI progress on Riemann zero bounds by Anthropic and special-case Hodge or Birch–Swinnerton-Dyer results from OpenAI releases has narrowed focus to those two problems, while lower odds on Yang-Mills, P versus NP, and Riemann reflect fewer credible demonstrations and greater foundational complexity. Competitive multi-agent workflows across labs, upcoming formal checks, and Clay Institute scrutiny remain key swing factors that could shift consensus before the 2027 cutoff.