Data Centers, AI, and the Grid: Can Load Flexibility Unlock New Capacity?

As artificial intelligence drives unprecedented growth in electricity demand, data centers are rapidly becoming some of the largest—and most consequential—loads on the U.S. power grid. Utilities that haven’t seen meaningful load growth in decades now face mounting interconnection backlogs, rising costs, and growing concerns about reliability, emissions, and equity.In this episode of People, Places, Planet, host Sebastian Duque Rios is joined by Dalia Patino-Echeverri of Duke University and Aroon Vijaykar of Emerald AI to explore whether load flexibility offers a way forward. They examine how data centers and AI stress today’s grid, how modest and carefully designed curtailment could unlock significant new capacity without overbuilding infrastructure, and what emerging technologies and policies—from flexible interconnection to software-driven demand response—could mean for electricity affordability, grid reliability, and the future of AI development in the United States.The Driving Forces Behind a New Wave of Electricity Demand (2:12)What's Constraining the Grid? (6:18)Rethinking Grid Limits through Load Flexibility (17:20)Inside a Flexible Data Center (40:13)What This Means for Policy, Costs, and Emissions (54:13)Learn more by reading about Emerald AI's pilot in Phoenix and Duke's report on load growth and flexibility, Rethinking Load Growth: Assessing the Potential for Integration of Large Flexible Loads in US Power Systems. ★ Support this podcast ★