Artificial intelligence is transforming the electricity system from both sides of the meter simultaneously: as an optimization tool for system operations, including renewable integration and grid resilience, and as one of the most intensive categories of electricity demand in the history of the sector. This article examines the regulatory challenges that the rapid growth of AI-driven data center demand is generating for power system governance, drawing on recent regulatory decisions, technical documentation, and policy proposals from North America, Europe, and Australia.
The analysis is organized around four substantive dimensions — long-term predictability, operational reliability, economic affordability, and environmental sustainability — that trace a spectrum from the primarily technical to the irreducibly political. A fifth dimension, regulatory independence, cuts across all four: not a separate domain of analysis, but the institutional condition that determines whether (and which) governance of the other four is possible at all.
This paper by Mr. Luca Lo Schiavo, ERRA Regulatory Specialist, is the winner of the ICER Distinguished Scholar Award ("Next Practices" category), in connection with the XI World Forum of Energy Regulation (WFER), Tbilisi, 2026.
