Series are multi-part explorations of a single AI-energy topic. Each series draws on the author’s deep domain expertise to connect technology, infrastructure, markets, regulation, and strategy into a clear framework for understanding the AI era.
The U.S. grid has ~100 GW of hidden capacity that data centers could unlock with tiny, well-timed curtailments. With stronger price signals—geometric demand charges and TOU surcharges—flexibility becomes profitable, enabling rapid AI load growth without new generation.
Michael Leifman
AI is transforming utility operations as COOs face soaring electrification, data-center load, extreme weather, and aging infrastructure. The shift from reactive to predictive, AI-driven management is now essential for reliability, resilience, and meeting rising customer expectations.
Hannah Kaplan
AI’s energy problem isn’t shortage—it’s misalignment. GPU clusters run at just 60–70% utilization due to data bottlenecks, creating hidden flexibility. With minimal peak curtailment, the grid could integrate ~100 GW of new load. Smarter governance—not more power—is the real solution.
Michael Leifman
AI developers are inflating power demand with phantom data centers—projects that exist on paper but never connect. Utilities build for the illusion, regulators approve the excess, and ratepayers bear the cost. Increased transparency can break the cycle.
Michael Leifman
AI has broken the grid’s forecasting logic. Data-center demand could double—or stall—by 2028. The problem isn’t prediction, it’s governance. Smarter rules, transparent data, and flexible pricing—not overbuild—will determine how well the grid survives the AI era.
Michael Leifman