The Trump administration, through Executive Order 14409 and a subsequent secret 'voluntary framework' completed in August 2026, has created a de facto licensing regime for frontier AI models. The framework defines covered models using a classified benchmark, requires pre-release government access for up to 30 days, and restricts initial distribution to government-approved 'trusted partners.' The framework's text remains secret, and the administration has already used unclear authorities to force Anthropic to withdraw models and limit OpenAI's release. For financial institutions, this creates regulatory uncertainty and potential revenue volatility for AI vendors dependent on U.S. government approval for model releases.
409 other tracked developments touching the same claims.