Across the U.S., utilities are developing resilience plans, either in response to state or federal mandates, or through voluntary, proactive efforts. It should come as no surprise that on the heels of a major storm or wildfire, regulators respond with new requirements, and this will only increase as the risks of major outages caused by extreme weather and wildfire grow.
As utilities explore resilience solutions, they are increasingly finding that artificial intelligence lies at the heart of both entirely new products and enhanced versions of legacy products; AI can also be a way for utilities to develop their own approaches in-house.
Some utilities are at early stages, while others are further along. But for the industry as a whole, 2025 and 2026 will be critical years for establishing true industry frameworks for resilience, and for moving innovations from pilots to deployments at scale. It is the moment for the industry to move from reactive to proactive approaches to resilience.
Our upcoming conference, Power Resilience Forum ‘26, in partnership with The Ad Hoc Group, will be bringing together utilities, regulators, and innovators to move the conversation forward. Stay tuned for more as we develop the program and add new partners. And in the meantime, a walkthrough of the power resilience and AI landscape.
The states pushing change
After the devastating Winter Storm Uri in 2021, both Texas law, specifically SB 3, and Public Utility Commission of Texas rules require utilities to file weatherization and grid resilience plans, conduct regular inspections, and demonstrate compliance.
CenterPoint Energy, for one, earlier this year submitted its enhanced systemwide resiliency plan as part of its Greater Houston Resiliency Initiative that aims to significantly reduce outages due to extreme weather.
Meanwhile, New York’s Public Service Commission requires utilities — Con Edison, National Grid, Central Hudson, as well as smaller ones — to submit detailed resilience and climate adaptation filings.
And the California Public Utilities Commission mandates wildfire mitigation and resilience plans for investor-owned utilities, which include PG&E, Southern California Edison, and San Diego Gas and Electric.
Beyond these big three states, New Jersey, Massachusetts, Maryland, Illinois, Oregon, Washington, Colorado, and Minnesota are all also developing requirements for utilities to assess climate risks and submit resilience plans, often tied to submissions of grid modernization or integrated resource plans. In Idaho, Nevada, and Wyoming, utility commissions are requiring wildfire mitigation plans for all utilities.
Likely a bit further out, Connecticut, Rhode Island, and Maine are actively studying regulatory frameworks for mandatory utility resilience plans. Florida has directed electric utilities to submit storm protection plans, which are a form of resilience planning, since 2019, though broader adaptation mandates may follow.
The AI-resilience nexus
Which brings us to AI, and how you’re going to see it underpin nearly every resilience solution coming to market in 2025.
Last year, Latitude Intelligence released the Utility AI Insight Report with a goal of capturing all the ways AI was being integrated into solutions across a vast array of use cases. We found that from generation and storage to T&D and the customer, resilience made up much of the value AI was currently delivering. And in the rare cases where utilities were developing their own AI applications in-house, resilience again was the primary driver and area where they could realize the quickest benefits.
Outages are costly and raise the ire of customers and regulators alike. AI’s appeal to utilities is often its ability to take massive data sets and quickly make sense of their implications for grid reliability. The use cases most often described today include the following.
Outage prediction and fault detection. AI-powered predictive analytics can forecast which grid assets are most at risk in upcoming storms, heatwaves, or wildfire conditions. With the aid of sensors, AI enables real-time anomaly detection on the grid, helping identify incipient faults or failing equipment before an outage occurs. The outcome is typically improved preventive maintenance, faster response times, and reduced downtime.
AI-enhanced weather forecasting. Solutions from startups and established tech companies use machine learning algorithms to analyze vast amounts of environmental, meteorological, and grid data. By learning from historical weather patterns, real-time sensor inputs, radar, satellite imagery, and even unconventional data sources (like social media or IoT sensors), AI can identify subtle trends or correlations missed by traditional models, produce more localized, accurate, and timely forecasts, and dynamically update predictions as new data streams in.
Storm and event impact modeling. Machine learning models analyze weather data, historical outage records, and asset health to predict the impact of severe weather events (e.g., hurricanes, ice storms, wildfires) on specific grid sections. This is often, though not always, achieved through digital twins. Utilities can pre-position crews and resources based on AI-driven impact forecasts. The values here are faster restoration, reduced customer impact, lower operational costs.
Vegetation management. AI processes imagery, LIDAR, and other data sources from traditional ground patrols, drones, and satellites to detect trees or branches encroaching on power lines, which are a leading cause of outages, especially during storms or fires. Utilities should benefit from targeted, efficient vegetation management, lower risk of weather-related outages, or wildfires. This is — not to make too easy a pun — the real low-hanging fruit. While some utilities have built their own tools, most continue to evaluate supplementing what they’ve built with advanced new technologies available in the market.
Wildfire risk mitigation. AI analyzes weather, fuel moisture, topography, and historical ignition patterns to calculate real-time fire risk near power lines. Utilities use this to make data-driven decisions about Public Safety Power Shutoffs and grid reconfiguration. Already in operation around the country, these solutions reduce wildfire ignition risk, improve communication with the public, and support more targeted shutoffs.
Grid asset health assessment. AI aggregates data from sensors (IoT), maintenance logs, and real-time monitoring devices to assess the health and failure risk of equipment like transformers, substations, and poles. This helps utilities optimize asset replacement schedules, lower failure rates, and build more resilient infrastructure.
Automated and self-healing grids. These are perhaps the furthest out, but movement toward automated grids is indicative of how an intelligent software layer can potentially be leveraged to dramatically reduce outage duration and scope. Common to these solutions, AI-driven control systems can detect and isolate faults automatically, rerouting power and restoring service to unaffected areas without human intervention.
Over the past year there has been a steady stream of announcements from vendors around these use cases, creating what is arguably a true market sector around AI-based resilience tech.
Traditional utility vendors have been busy. IBM rolled out an AI-driven weather operations center for utilities, Siemens launched its Grid Resilience-as-a-service, GE Vernova has enhanced its ADMS with AI features for rapid reconfiguration of grids in real-time, and Oracle has added AI-powered outage prediction to its NMS. EPRI’s Climate READi initiative continues to expand its collaboration around developing standards and shared AI tools for utilities to improve resilience and adaptation.
And startups are just as busy. We looked at over 50 startups in our first Utility AI report, and since then the number has continued to grow. This is by no means an easy market to enter. Utilities are understandably cautious around new technology adoption, but in use cases like vegetation management, wildfire detection and mitigation, infrastructure monitoring, dynamic line rating, and weather forecasting there are strong signals that utilities can gain real resilience value.
Not far behind will be AI solutions that enhance grid visibility and situational awareness — a critical step towards further DER integration and orchestration — and other solutions that sit with the customer and improve outage detection, communications and customer service.
A version of this story was published in the AI-Energy Nexus newsletter on April 23. Subscribe to get pieces like this — plus expert analysis, original reporting, and curated resources — in your inbox every Wednesday.
