Mediu

January 9, 2026, 4:00 AM EST. MIT's Priya Donti explains how AI could sharpen power-grid optimization. The aim is to keep supply and demand in balance in real time amid uncertainty on demand, fuel costs, and weather-driven renewable output. AI can fuse historical and real-time data to forecast how much renewable energy will be available and to solve complex trade-offs: which generators to run, how much to produce, when to charge or discharge batteries, and how to flex loads. These optimization tasks are computationally heavy, so operators rely on approximations that can be inaccurate, a problem as renewables grow. Better AI methods could push toward a cleaner, more reliable grid, though challenges remain in scaling, risk, and real-world deployment.