AI on the Grid: How Predictive Analytics Aims to Stop California Wildfires

LIBERTY LAKE, WA – February 03, 2026 – In the ongoing battle against California's devastating wildfires, a new technological front has opened on the state's sprawling electrical grid. San Diego Gas & Electric (SDG&E) has teamed up with technology firms Itron and Toumetis to deploy a sophisticated system of predictive analytics and artificial intelligence, aiming to stop grid failures before they can spark a blaze or plunge communities into darkness.

The project, operating across more than 800 electric distribution circuits in Southern California, represents a significant shift from a reactive to a proactive model of grid management. By identifying the subtle warning signs of equipment failure in real-time, the collaboration seeks to enhance public safety in high-fire-threat areas and dramatically improve electric reliability for millions of customers.

A New Paradigm in Grid Intelligence

At the heart of the initiative is the fusion of two powerful technologies: Itron’s distributed intelligence-enabled endpoints and Toumetis’ Cascadence™ predictive analytics platform. Itron, a leader in intelligent infrastructure, provides a constant stream of high-resolution data from its network of advanced smart meters and grid sensors. This granular data, which includes detailed electrical waveforms, offers an unprecedented view into the health of the grid at its very edge.

This torrent of information is then processed by the Cascadence™ platform, which uses what Toumetis calls “physics-informed analytics.” This approach combines advanced machine learning algorithms with the fundamental laws of physics that govern electricity. By embedding expert knowledge of how electrical systems should behave—and how they behave when they are about to fail—the AI can more accurately distinguish between normal fluctuations and genuine precursors to an outage. The system unifies these signals with data from substation power quality monitors and other sensors into a single, cohesive operational view.

“Reducing SAIDI requires both prevention and faster restoration,” said Mark Willnerd, Chief Executive Officer of Toumetis Inc., in a statement. “Cascadence is built to unify disparate grid data, detect subtle precursors to failure and deliver actionable intelligence in minutes.”

This integrated system is designed to detect emerging risk conditions far earlier than traditional methods, enabling the utility to prioritize the most critical threats and dispatch crews for targeted intervention.

From Reactive to Proactive: Preventing Outages Before They Start

The primary goal of this advanced system is to prevent outages from ever occurring. Historically, utilities have often operated in a reactive mode, scrambling to locate and repair faults only after they cause a customer-impacting power interruption. This new approach enables SDG&E to identify early warning signs—such as abnormal voltage behavior, repeating waveform patterns that indicate equipment stress, or the subtle signatures of failing components—days or even weeks in advance.

When the Cascadence™ platform flags a potential issue, it provides actionable intelligence that allows the utility to address the problem proactively. This could involve dispatching a crew to replace a stressed transformer or reinforce a weakened connection before it fails. By fixing problems before they escalate, the project directly targets a reduction in the System Average Interruption Frequency Index (SAIFI), a key industry metric that measures how often the average customer experiences an outage.

“We have a strategic relationship with Toumetis to incorporate Cascadence with Itron’s waveform detection and classification agents on Gen5 Riva meters to provide grid edge analytics and operational awareness,” noted Don Reeves, Senior Vice President of Outcomes at Itron. This collaboration, he explained, helps utilities “detect and locate precursor conditions, prioritize detection risk and restore power faster.”

The Race Against Time: Accelerating Restoration and Mitigating Fire Risk

While prevention is the ideal, the partnership also aims to drastically shorten the duration of outages that do occur. When a fault happens, finding its exact location on miles of power lines can be a time-consuming process involving extensive patrols by field crews. This delay extends restoration times and, in high-risk areas, increases the window for a downed power line to ignite a fire.

The Itron-Toumetis system addresses this challenge by using its physics-based location algorithm to pinpoint the source of a fault with far greater precision. By providing crews with an accurate location in minutes, the technology significantly reduces patrol time and accelerates dispatch and repair decisions. This directly tackles the System Average Interruption Duration Index (SAIDI), which measures the average length of outages.

This speed is particularly critical for wildfire mitigation. In Southern California's dry, windy conditions, a spark from failing electrical equipment can rapidly escalate into a major catastrophe. By prioritizing potential ignition-related electrical conditions for immediate investigation, the system provides a crucial tool for enhancing safety in designated high-fire-threat districts, allowing SDG&E to respond before a potential ignition source can cause harm.

A Multi-Billion Dollar Effort Backed by State Innovation

This high-tech initiative is not happening in a vacuum. It is the latest step in a massive, long-term effort by SDG&E to harden its infrastructure and prevent a repeat of the devastating 2007 wildfires, some of which were caused by its equipment. Since then, the utility has invested approximately $6 billion in ratepayer funds into a comprehensive wildfire mitigation strategy, replacing thousands of wood poles with steel and adopting an aggressive ethos of public safety.

The project is also supported by California's Electric Program Investment Charge (EPIC) program. Funded by a small surcharge on customer bills, EPIC is a state-administered fund designed to drive innovation and support the demonstration of pre-commercial clean energy and grid modernization technologies. This regulatory framework creates an environment where utilities can test and deploy cutting-edge solutions like the Cascadence™ platform, which might otherwise be considered too novel for immediate, large-scale investment.

The results of this initial deployment will be closely monitored, with metrics tracking prevented events, time-to-locate, and time-to-restore. Positive outcomes are expected to guide a potential expansion of the program into additional circuits and operating scenarios, further solidifying the role of AI in the utility's operational toolkit.

As climate change continues to amplify wildfire risk, the integration of predictive analytics into grid management represents a critical evolution. This partnership between Itron, Toumetis, and SDG&E is a clear signal that the future of utility safety and reliability lies not just in stronger poles and wires, but in the intelligent, predictive power of data.