Morgan Bazilian, director of the Payne Institute for Public Policy, argues for verified large-load governance of AI, including milestone-based forecasting, cost-causation tariffs, flexible service classes, operational visibility, and coordinated federal-state oversight.