Contributing authors:
Hari Vasudevan (PE), Founder and CEO of KYRO AI
Greg Smith, Senior Manager, Vegetation ROW Management at LS Power Grid
Eric Easton (PhD, PE), VP Grid Transformation & Investment Strategy, CenterPoint Energy
It’s time to hedge in vegetation. Power outages and disasters from overgrown vegetation are impacting the grid’s reliability and America’s energy security. The challenges faced by electric utilities have a direct impact on the power grid’s resiliency and on the energy affordability many Americans grapple with today. A 2019 survey of North American utilities found that about 23% of power outages were attributable to overgrown vegetation. Yet barriers exist to solving these challenges, including limited manpower, large land areas to monitor, and budget constraints. Taking action before vegetation becomes hazardous is of paramount importance. It’s time to shift from reactive trimming cycles to proactive vegetation management, combining Artificial Intelligence (AI), lidar, and geospatial data to do so.
Utilities and Vegetation Management Challenges
Cycle-based approaches are largely the norm in vegetation management, but these approaches only treat vegetation once it’s already hazardous. That drives up costs and leaves communities and utilities vulnerable to risk. To illustrate, in recent years, fires in California, Texas, and Hawaii have caused loss of life, catastrophic property damage, and prolonged interruptions of utility services to critical infrastructure. Federal, state, local, and private lawsuits have been brought against public and investor-owned utility companies. Many of these events are tied to physical aspects of maintenance and the related vegetation on these systems. Vegetation-fueled fires and grid equipment failures pose existential legal and financial risks, underlining the need for more proactive vegetation management.
The Financial Toll of Outdated Processes on Utilities
According to the U.S. Department of Energy (DOE), power outages cost about $150 billion every year, and vegetation management takes up a significant share. However, more than half of utilities admit their budget for vegetation management does not meet operational requirements. The industry also faces an aging workforce and growing difficulty in attracting skilled field and technical talent. Inspections often take multiple field days per year, with additional desk days for office staff for reviewing and reporting. Both field and office work require qualified field and technical talent. These challenges compound the problem of manually collecting and processing data. Utilities inspect hundreds of thousands of miles each year, producing mountains of unstructured data that are difficult to standardize, analyze, and act on manually. From a business intelligence standpoint, manually aggregating and acting on this data without AI becomes practically impossible.
From Manual Cycles to Data-Driven Decisions
The combined use of AI, lidar, and geospatial data closes the gap between field-generated data and instantly synthesized actionable insights. lidar and satellite-derived risk maps guide field crews. Together, these solutions turn field reports into a unified and proactive view to identify risks with greater efficiency. Field notes and photos are turned into real-time indicators for hazardous trees, ground-to-sky work, and access needs. The transformation from reactive to proactive operations is profound. Instead of relying on fixed trimming cycles every three to five years, utilities can now shift to risk-based schedules. Technology reduces blind spots, allows for increased frequency of surveys, and alleviates operational and labor barriers. Utilities can use AI risk models to target the most vulnerable corridors instead of trimming every mile of line on a fixed schedule, thereby reducing outage risk and wildfire exposure. Importantly, AI can also produce predictive treatment plans based on historical data in specific locations. This improves regulatory compliance and is much more cost-effective. The predictive capabilities particularly allow utilities to work smarter by focusing limited labor pools and budgets where manual interventions can have the greatest impact on reliability. Enhanced resource planning and allocation benefit crews so they work in areas that require manual attention, resulting in a much more manageable workload. Moreover, utilities that pair machine learning and image-based data can identify high-priority areas based on risk and predict fuse failures. They can also mitigate storm-inflicted damage by performing last-minute vegetation management in areas with predicted fuse failures. This level of risk management is critical to strengthening U.S. energy security and grid resilience in the long run. Additionally, modern offline field-capable data collection tools can be implemented in field inspections and tree removal activities. This information can supplement lidar and geospatial data with real-time images and measures that improve model awareness and leverage the presence of field resources acting on information produced by AI models and plans. Together, these solutions provide utilities with a holistic, efficient, and informed mode to manage vegetation.
Results from CenterPoint Energy and LS Power
To illustrate the real-world impact of these technologies, CenterPoint Energy has used lidar data across its transmission and distribution systems for two years. This has enabled advanced analytics and physics-based simulations for vegetation and damage prediction. By modeling how varying wind speeds and directions affect vegetation and equipment, CenterPoint Energy’s lidar-based digital twin pinpoints the most at-risk line sections and optimizes restoration priorities. Machine learning further assesses site accessibility to improve recovery efficiency, allowing CenterPoint to plan based on both risk and restoration difficulty. With annual reviews and multi-year data comparisons, CenterPoint Energy can track vegetation growth and infrastructure changes more precisely. While cost data is still being collected, the approach supported CenterPoint’s Greater Houston Resiliency Initiative and delivered a 50% year-over-year reduction in vegetation-related outage minutes in 2025, improving both reliability metrics and customer experience. At LS Power, years of vegetation performance data demonstrate the value of a predictive approach to vegetation management. After shifting to high-density ground-cover conversions and annual maintenance in targeted areas, the program is on track to achieve lower steady-state costs per acre across the entire right-of-way system. Total vegetation management spending over the coming decade-plus is projected to be lower than historical spending over a comparable prior period. This demonstrates the long-term cost savings and operational efficiencies achieved by implementing AI-driven vegetation management strategies.
Making the Business Case for Smarter Vegetation Management
While the cost benefits are compelling, there are key factors to consider before adopting and deploying these solutions. Convincing executives to invest in technology requires financial justification too. Fortunately, there is strong potential for achieving a return on investment. First, preventing a single wildfire or large-scale outage can save hundreds of millions, even billions, in potential liabilities and fines. One DOE-cited case hones in on this: A Midwestern electric cooperative reported a 30% cut in tree-related outages and a 45% reduction in outage duration by tracking and analyzing vegetation impacts using technology. These improvements were achieved while trimming costs fell by about 80%. Operational costs can also be significantly reduced thanks to lidar and AI analysis, which handles the workload in weeks that it would take crews months to do. Review times are slashed via the automated interpretation of field notes and photos. Switching to targeted risk-based scheduling saves money by trimming only where needed. More importantly, executives want the burden of proof quantified before they commit to this investment. CenterPoint Energy is saving millions of customer minutes of interruption annually by predicting outages with more than 80% accuracy. LS Power expects material long-term financial savings in vegetation management-related activities. The final ROI dimension is the regulatory one. The North American Electric Reliability Corporation (NERC) vegetation management standards require specific documentation. AI enables utilities to capture and organize time-stamped, location-specific clearance records that stand up to NERC’s compliance reviews and audits. These savings will likely accelerate as AI models mature and data volumes grow, creating a virtuous cycle. Better resource use improves rights-of-way (ROW) data quality and coverage, which sharpens AI models and enables smarter, more efficient budgeting without sacrificing momentum or manpower. AI has the potential to connect the most relevant information to utilities’ limited personnel, thus focusing efforts, budgets, resources, and priorities on the areas affecting reliability. It’s an important bridge between the cost of collecting much-needed information and executing critical aspects of vegetation management. These tools give utilities a past, present, and predictive view for better decision-making and operational outcomes. Together, these advances will help deliver a more resilient, reliable, affordable, and secure American electric grid.
