OpenAI’s September 8 announcement of an internal multi-agent system solving the Navier-Stokes Millennium Prize Problem in 88 hours with a Lean-verified proof has reset expectations for further breakthroughs. That result, produced amid concurrent human work by researchers at NYU and Anthropic, demonstrated frontier-model scale on one of the seven Clay problems without the company claiming the prize. Traders are now watching October’s release of over 700 additional math papers, which included progress on related conjectures, as evidence of accelerating capability. Key swing factors include independent verification timelines, potential credit or priority disputes, and whether OpenAI’s next internal model yields another formally checked Millennium solution before year-end deadlines.