OpenAI’s early October release of 722 AI-generated manuscripts claiming solutions or advances on hundreds of open problems, including partial results tied to the Hodge Conjecture, quasi-Riemann hypothesis, and Unique Games Conjecture, has become the dominant driver of trader sentiment. Many proofs include Lean formalizations, yet most await independent peer review, creating uncertainty over whether qualifying announcements will meet market criteria by year-end. Smaller labs such as doubleAI have also delivered verified progress on long-standing graph-theory questions at low cost, while competitive pressure among frontier labs accelerates further releases. Key catalysts ahead include verification outcomes, additional model announcements, and any formal claims on Millennium Prize problems before December 31.