AI could make power grid more efficient — if utilities can persuade regulators

By Zack Colman | 05/18/2026 06:51 AM EDT

Advocates say a yearslong pattern of rejection from regulators has chilled the pursuit of those new tools — even as power companies are struggling to keep up with the demand from their massive data centers.

One of Pacific Gas and Electric's Diablo Canyon Power Plant's nuclear reactors in Avila Beach, California.

One of Pacific Gas and Electric's Diablo Canyon Power Plant's nuclear reactors in Avila Beach, California, on Nov. 3, 2008. Michael A. Mariant/AP

If the burgeoning crop of new technology tools for wildfire prevention were designed for anyone, Pacific Gas & Electric would be the poster child.

Nearly six years removed from bankruptcy after a settlement for its role in the catastrophic 2018 Camp Fire in California, the Oakland-based utility is pursuing a round-the-clock monitoring program for power outages and wildfires. It pulls readings from 5.5 million meters that flag compromised infrastructure, 1,600 weather monitoring stations that use artificial intelligence to forecast fire-risk conditions and 600 infrared cameras trained on faint flickers to detect wildfires across its service territory.

But PG&E is facing resistance to the needed spending on the program from state regulators who are focused on keeping costs under control for residents and businesses.

Advertisement

The advent of AI-supported software and hardware offers electric utilities powerful new technology to more efficiently manage power resources, extend the lifelines of expensive equipment and quickly respond to costly power outages. But the strain of rising power prices has created hurdles that are often keeping those innovative models out of the hands of electricity providers.

GET FULL ACCESS