The Trump administration faces a self-imposed Aug. 1 deadline to unveil a framework implementing its June AI executive order, as Big Tech intensifies lobbying over how far Washington should go in policing advanced models. The plan is expected to rely on voluntary model submissions and a classified benchmarking process to determine “covered frontier models” and designate “trusted partners” with access. The countdown comes amid a sharp industry split over whether to restrict Chinese “open-weight” systems, with Nvidia CEO Jensen Huang urging policymakers to avoid “premature restrictions” and a slate of tech firms backing that stance. OpenAI signaled broad support while Anthropic outlined a separate position; both companies have recently navigated ad hoc limits, including export-control pauses and constrained rollouts at the government’s request. The framework could clarify a murky deployment environment, but the fight over open-weight access—and who shapes the rules—signals more regulatory battles ahead.




























