New AI Plan Keeps Your Lights On by Moving Batteries Around

Two researchers, Ali Abbasi and Kyri Baker, published a plan for keeping the lights on when storms, wildfires, or heat waves knock out power lines. Their idea centers on what they call Mobile Energy Storage Systems — basically giant batteries on wheels, hauled by truck to wherever they are needed. Right now, utilities mostly rely on fixed equipment. Rolling batteries in when trouble hits is faster and more flexible.

The plan works in three phases. On a normal sunny day, the batteries sit at power stations and make money: they charge when electricity is cheap and sell it back when prices spike, a practice called arbitrage. That's like buying gas on a Tuesday because it's cheaper. When forecasters warn a storm is coming, phase two kicks in. The AI picks the best parking spots — near hospitals, shelters, and busy neighborhoods — using weather models that predict which power lines are most likely to fail. It's essentially choosing the best staging ground, like moving ambulances closer to where accidents are expected. After the storm hits, phase three keeps re-routing the batteries as new outages appear.

To make these decisions quickly, the team built a graph neural network — an AI that understands how things are connected, in this case the network of power lines and roads. Without it, the math would be too slow to be useful during a fast-moving emergency. Their simulations showed that this approach reduces the total electricity customers go without and speeds up restoration.

The catch: this is a computer simulation, not a real utility pilot. Real-world deployment needs utilities to buy mobile batteries, which cost millions each, plus coordination with truck fleets and regulators. The paper doesn't address who pays or how many batteries a city would need. Still, as extreme weather gets more common, the question of how to keep the lights on is becoming a mainstream one for every household.