Can AI revolutionize grid operations?

This is a look at how utilities plan and run the electric grid, and how artificial intelligence might change that work behind the scenes. Josh Wong, founder of ThinkLabs AI, explains that utility planning today relies on slow, manual studies that check for worst-case scenarios, like the hottest day of the year, and often take months to run. When problems turn up, the fix is usually more wires and equipment, since planners rarely have tools that suggest cheaper alternatives like batteries or shifting demand.

Wong's company builds AI trained on the physics of the grid rather than general-purpose chatbots, aiming to cut study times from months to minutes and suggest actual fixes, not just flag problems. He argues this same approach could help grid operators too, by learning from daily events, catching patterns in alarms, and connecting long-term planning decisions to real-time operations, which today barely talk to each other.

The conversation also touches on data centers building their own gas generators and batteries to get power faster instead of waiting on grid connections. Wong is skeptical this is the best path: he believes the grid already has enough spare capacity for most data centers if utilities planned and operated it better, and he suggests large data centers should be treated less like private backup systems and more like their own small utilities with responsibilities to the surrounding grid.

For homeowners, none of this changes anything immediately. It is a preview of technology utilities may adopt to plan and manage the grid more efficiently, which over time could affect reliability and how new demand, like data centers, gets connected.