OpenAI and Anthropic's late-October 2026 announcements that internal large language models independently produced subexponential-time algorithms for #3-SAT and k-CNF counting, directly refuting the Exponential Time Hypothesis, represent the core catalyst shifting trader sentiment. Public GitHub releases included natural-language papers and Lean formalizations, prompting an expert review group led by ETH originators Russell Impagliazzo, Ramamohan Paturi, and Ryan Williams to examine the claims alongside hundreds of papers relying on the conjecture. While the results remain unpeer-reviewed and show crossover points too large for practical impact, they highlight AI-driven progress in complexity theory and competitive dynamics between frontier labs. Key upcoming catalysts include the group's verification timeline, potential conference discussions, or formal community consensus that could resolve confirmation markets.