Anthropic and OpenAI have driven recent trader sentiment through September commitments to embed third-party evaluators with employee-level access for safety assessments, including Anthropic's contract with Accenture's Faculty unit and OpenAI's ongoing contract finalization with groups like METR. A late-September White House voluntary accord, signed by executives from OpenAI, Anthropic, Google, Meta, Nvidia, and xAI, explicitly encourages independent external auditors to review internal controls, with President Trump voicing support for industry self-policing. Key uncertainties include evaluator funding, publication rights for findings, and the precise scope of access, while California’s new auditor registry and the proposed FRONTIER Act add regulatory context. With the October 31 deadline weeks away, traders are watching for formal announcements or expanded sign-ons that could clarify which labs meet the vetting threshold.