PG&E's new center uses machine learning to prevent wildfires
Pacific Gas and Electric Co. launched a new monitoring center that uses machine learning to help detect and prevent wildfires.
The Continuous Monitoring Center in San Ramon utilizes data from tens of thousands of sensors and the company’s smart meter network to identify potential issues before they escalate, PG&E said in a community announcement.
The center analyzes information from approximately 5.5 million meters to detect abnormalities in the electrical grid. This allows experts to dispatch crews to address problems before they lead to wildfires or outages, PG&E spokesperson Paul Moreno said.
PG&E: Center effective in preventing fire starts, reducing outages
In 2025, the monitoring system intercepted 17 potential ignitions in high fire-risk areas, avoided 12 million minutes of unplanned customer outages, reduced emergency outage response time by 2,620 hours and saved approximately $6 million in operational costs, the utility said.
The center brings together various technologies, including early fault detection sensors, GridScope devices, downed conductor detection and SmartDetect, which uses machine learning to monitor grid performance. These tools work together to identify and address hazards earlier and more precisely, Moreno said.
The Continuous Monitoring Center is part of PG&E’s broader wildfire mitigation efforts, which include undergrounding, Enhanced Powerline Safety Settings, Public Safety Power Shutoffs, AI-enabled wildfire cameras, according to the company.
Note to readers: If you appreciate the work we do here at the Redding Record Searchlight, please consider subscribing yourself or giving the gift of a subscription to someone you know.
The new system marks a shift from reactive response to proactive detection and prevention, Moreno said, especially as wildfire season begins.
The center’s team of engineers identified a wiring issue on the Brunswick 1106 circuit in Nevada County. The machine learning model detected a possible problem, and an electric troubleshooter found melted insulation at a transformer caused by severe weather-related stress. Crews replaced the transformer and other equipment before an ignition could occur, Moreno said.
Since 2025, PG&E has recorded 1,484 similar catches. The center analyzes data from multiple technologies deployed across PG&E’s systems, including early fault detection sensors, GridScope devices, downed conductor detection, SmartDetect, distribution fault anticipation sensors, line sensors and a grid data analytics and technology platform.
For more information, go to pge.com/news.
This story was created by Jessica Skropanic, jessica.skropanic@redding.com, with the assistance of Artificial Intelligence (AI). Journalists were involved in every step of the information gathering, review, editing and publishing process. Learn more at cm.usatoday.com/ethical-conduct.
This article originally appeared on Redding Record Searchlight: PG&E using AI to prevent wildfires and outages