OpenAI, Anthropic and Google DeepMind are pursuing an industry-led AI safety standards organization, with discussions underway since July 2026 and regular working-group meetings by mid-September. The proposed framework would cover shared technical evaluations, pre-release audits, independent testing and standardized safety protocols. Google DeepMind CEO Demis Hassabis proposed a U.S.-led body modeled on FINRA on July 14, while OpenAI chief scientist Jakub Pachocki called for common sector-wide requirements in a September 6 essay. The effort builds on the 2024 Frontier Model Forum but excludes Microsoft and follows the companies’ decision not to join Nvidia’s Open Secure AI Alliance. The initiative could shape enterprise contracts, government procurement and public trust. Separately, prediction markets give Anthropic a 96% chance of producing the best AI model by September 2026, while higher valuation thresholds have gained support, though these remain market expectations.