Anthropic CEO Dario Amodei’s September 12 proposal for permanent, employee-level access by independent evaluators to frontier AI models and training processes, swiftly endorsed by OpenAI’s Sam Altman, has driven recent trader sentiment in this market. OpenAI reinforced the stance with a September 22 framework detailing priority areas for safety-case assessments across training, evaluation, and deployment, plus principles for access, independence, and publication rights. These moves reflect competitive dynamics among leading labs amid growing AI risk concerns, though Google, Meta, Microsoft, and xAI have yet to match the commitments, keeping their market-implied probabilities low. With the October 31 deadline approaching, further announcements or legislative signals like the FRONTIER Act could shift outcomes before resolution.