For entrepreneurs with a provocative vision for energy & IT; NGP is the only utility venture & innovation fund that taps into the global network of National Grid, accelerating companies with investment, innovation, incubation & biz dev; realizing the full potential of energy.

National Grid Partners

Electric utilities today face mounting challenges – from explosive load growth to aging infrastructure. At a recent NextGrid Alliance (NGA) working group session on AI-powered grid operations, utility leaders and startups innovators shared how they’re using data-driven tools to move the grid from reactive firefighting to forward-looking planning.  

National Grid Partners founded the NGA in 2020 to help like-minded utilities tackle shared challenges through technology and collective intelligence. Today, the group includes more than 500 senior professionals from 135 utility companies in 25 countries.  

The NGA’s Grid Modernization Working Group co-chair John Franklin reminded session participants that utility infrastructure is critically important for customers and for society. "Data underpins every one of those responsibilities,” said Franklin, Vice President of Asset Strategy at National Grid.

AI-Powered Grid Operations: From Reactive Repairs to Predictive Reliability

Franklin’s team explained how artificial intelligence can flag which electricity distribution circuits are most at risk of poor performance. In Massachusetts, for instance, utilities must report circuit performance with a focus on ensuring no customers experience subpar reliability, and they face penalties if a circuit repeats bottom performance YOY. There are many ways the company approaches the remediation of these “poor performing circuits”, but AI is changing this and allowing data to help guide the areas of focus to get remediation work started earlier than ever before. 

National Grid IT’s data science team builds Monte Carlo–style simulations to predict which feeders will land in the bottom tier of reliability scores by year’s end. The model looks at each feeder’s unique attributes (length, forest coverage, last tree trim, etc.) and runs hundreds of potential scenarios. Armed with forecasts of the bottom performers, Franklin’s team gains weeks of lead time to intervene; instead of waiting for outages to pile up, planners can target upgrades and vegetation work where the AI model predicts failures.  
This shift to a more proactive one reduces the chance of repeat poor reliability performance and improves customer satisfaction.  

And over time, even first-year problem circuits can be treated before they become repeat offenders. With predictive analytics, combining diverse datasets (e.g. weather, maintenance history, equipment age, etc.) can reveal insights no single spreadsheet could. 

Predictive Feeder Analytics at National Grid

Southern California Edison (SCE) is pursuing digital transformation on multiple fronts. Ray Fugere, Director of Asset and System Intelligence at SCE, described how his team is creating a “digital twin” of the grid – a 3D model of every pole, switch, and line – using LiDAR surveys. This digital terrain lets engineers visualize asset locations and overlay real-time weather or loading data. As Fugere explained, SCE has already “painted” its service territory with LiDAR to build a virtual replica of the electrical network. The digital twin will allow system planners to simulate different grid conditions and make smarter investment decisions for planning and outages. 
 
Computer vision is another area of innovation at SCE. Fugere highlighted an AI project to audit switchgear inventory using imagery. The team built an in-house model using the YOLO (You Only Look Once) object-detection framework to spot switches on utility poles and classify their type. This model, Fugere said, now achieves “about 88% accuracy in classifying the type of switch” in a photo. That level of detail can help update asset registers and direct crews to where hard-to-find equipment is installed. 
 
Together, these digital tools empower operators with unprecedented visibility. As an example, SCE develops Advanced Waveform Anomaly Recognition System (AWARE), which leverages AI and machine learning to predict and identify potential abnormalities on the grid that the human eye may overlook. By 2030, Fugere expects the use of even more AI tools to be “normal” – from automating data entry by embedding language models in work-order systems, to using computer vision for quality checks on new construction. Fugere even foresees large language models sifting through lineworker field notes to pinpoint failure causes, and AI agents continuously curating grid models as weather conditions change. 

Building a “Digital Twin” and Leveraging Computer Vision

Another perennial reliability issue is vegetation growth under power lines, and both SCE and National Grid are experimenting with artificial intelligence to tame it.  

Fugere noted that SCE spends roughly $500–700 million annually on tree trimming and vegetation-related work. To optimize this, the utility is partnering with satellite-data startups to provide continuous monitoring. By analyzing periodic satellite imagery with AI, these startups can detect tree growth near lines and even estimate growth rates year over year. This means instead of dispatching crews on fixed schedules, foresters can prioritize only the sections where trees are encroaching.  

“Using satellite imagery with AI to detect vegetation growing,” Fugere explained, helps SCE make “smarter decisions on our vegetation budget” and trim trees more strategically.  Companies like AiDASH, a National Grid Partners portfolio startup, offer similar vegetation monitoring services using satellite and aerial data. In Massachusetts, National Grid has seen storm-related outages fall by 30 percent since deploying AiDASH technology in 2022 to help crews identify and address vegetation growth. 

AI-Driven Vegetation Management

Electricity forecasting is yet another AI application making waves. Sean Kelly, co-founder and CEO of Houston startup Amperon, noted that utilities globally face up to 35–50% growth in electricity demand by 2040 due to trends like the rise of data centers and EVs. Amperon feeds data from hundreds of sources (including weather, meters and system operators) into AI models to produce minute-by-minute forecasts of energy demand, solar and wind output, and prices. These forecasts help utilities plan generation, procurement, and grid upgrades more efficiently. (Indeed, National Grid Partners recently backed Amperon as part of its AI portfolio, recognizing the need for data-driven demand planning.)

Forecasting Future Demand

John Franklin’s closing thought encapsulated the spirit of the session: Every ounce of infrastructure reliability comes down to making data work smarter. By 2030, he and others envision artificial intelligence being part of the grid’s DNA. As one attendee put it, utilities are moving past “will we use analytics?” to “how will we do this?” 

Vision for 2030: Smarter, Seamless Operations

AI is enabling utilities to transition from reactive fixes to predictive planning. Real examples include: 

Predicting worst-performing feeders with Monte Carlo simulations to prioritize upgrades and reduce reliability penalties. 

Using YOLO-based computer vision models to automatically inventory and classify grid assets with ~88% accuracy. 

Creating a full 3D grid “digital twin” via LiDAR scans to simulate performance and guide smarter infrastructure planning. 

Monitoring vegetation growth with satellite imagery and AI to optimize trimming cycles 

Deploying advanced waveform analytics to detect grid anomalies, leveraging AI/ML to identify issues the human eye might miss and improve fault prediction accuracy. 

Startups backed by National Grid Partners are part of this story, including Amperon (AI-driven demand forecasting) and AiDASH (AI in vegetation management).

Key Takeaways

Shun Sakaguchi is a Manager of Innovation & Venture Acceleration at National Grid Partners, the venture investment and innovation arm of National Grid. He leads the NextGrid Alliance, which gives utilities and other industry stakeholders front-row access to technology advances. To stay informed on the latest in AI and other grid modernization topics, follow National Grid Partners or request to join the NextGrid Alliance.  

By: Shun Sakaguchi, National Grid Partners