Electric systems are being asked to serve loads they were never designed for. Electrification, data centers and shifting customer behavior are pushing peaks higher and straining infrastructure-built decades ago. One of the fastest-growing contributors to that strain, the electric vehicle (EV), is also one of the most valuable tools a utility has to manage it. The difference between a problem and an asset comes down to one question: can you see it?
EVs are no longer a future planning assumption. They are on distribution systems today, in growing numbers and they charge in ways that matter to operations. A Level 2 home charger can pull more than six kilowatts, roughly the peak demand of an entire house and it usually switches on in the early evening, exactly when the system is most stressed. Because cars sit parked most of the day, that load is also far more flexible than air conditioning or water heating, which customers feel immediately. Time-of-use (TOU) rates typically move 60 to 70 percent of EV charging to off-peak hours and managed charging can shift more than 90 percent. At scale, that flexibility can shave system peaks and defer distribution upgrades when both are expensive.
So why are so few utilities capturing it? Because almost none of that flexible load is actually being managed. In a 2026 National Bureau of Economic Research study, researchers ran a managed-charging experiment across EV households at a Bay Area utility with one of the highest EV adoption rates in the country. Even when households were paid $40 a month to take part, fewer than five percent enrolled. The broader picture isn’t any better: the Smart Electric Power Alliance reports that where residential TOU rates exist at all, opt-in adoption rarely tops 10 percent. Compound the two and most utilities are actively managing only a small fraction, often in the single digits, of the EV load already on their systems. Voluntary sign-up is not a path to grid visibility.
Knowing the opportunity is real is not the same as knowing where it lives. To act, a utility needs to know how many EVs are on its system and where and when they charge. This is where most efforts stall. Vehicle registration data is coarse and outdated. Enrollment and rebate records capture only the customers who raised their hands, while the drivers running unmanaged chargers that strain a local transformer are often the least likely to do so. The EVs that pose the greatest risk are often the ones the utility cannot see.
The answer is already in the utility's own data. Years of 15-minute interval readings from Advanced Metering Infrastructure (AMI) carry a distinct signature every time an EV charges. The challenge has always been reading that signature reliably across hundreds of thousands of meters. That is the problem NewGen built GridLens™ to solve.
GridLens™ applies the same family of pattern-recognition technology (convolutional neural networks) that powers medical imaging and satellite analysis, reading ordinary consumption data to find the visual fingerprint of EV charging. It flags likely EV charging with enough confidence to stand behind real decisions, from distribution planning to program targets to rate design. The result is a defensible, probabilistic map of where EVs charge, built entirely from data the utility already owns.
Just as important is who can use it. GridLens™ runs through a simple interface built for planning and rate staff, not data scientists. A utility points the tool at its own AMI data and receives results in minutes, not months. The tool does the heavy lifting, normalizing AMI data and adjusting automatically for differences across climate zones and rate classes.
Visibility is the starting point, not the finish line. Once a utility can see its EV load, it can act based on evidence rather than assumptions: target the right circuits, design rates customers will actually respond to and track whether those programs work. Left unseen, EV load quietly becomes a planning problem. Seen clearly, it becomes the flexible capacity utilities are already paying to find elsewhere.
AMI was sold on a promise: that all this data would make the grid smarter, easier to plan and more efficient, not just produce a more accurate bill. EV detection is one example of that promise finally being delivered and it is not the last. The same approach can surface other loads hiding in the same data. The capacity many utilities are seeking is already moving through their systems every 15 minutes. Now we have a way to see it.
Sources:
Burlig, F., Bushnell, J. B., & Rapson, D. S. (2026). If You Build It, They May Not Come: Willingness to Participate in Managed EV Charging. NBER Working Paper No. 35086. https://www.nber.org/papers/w35086
Smart Electric Power Alliance (SEPA). Set It and Save: Using Technology to Ensure a Smooth TOU Rate Transition. https://sepapower.org/knowledge/set-it-and-save-using-technology-to-ensure-a-smooth-tou-rate-transition/
Bailey, M. R., Brown, D. P., Shaffer, B. C., & Wolak, F. A. (2023). Show Me the Money! Incentives and Nudges to Shift Electric Vehicle Charge Timing. NBER Working Paper No. 31630. https://www.nber.org/digest/202401/shifting-electric-vehicle-owners-peak-charging
ev.energy. Dynamic Pricing Outperforms Time-of-Use in California EV Charging Pilot. https://www.ev.energy/blog/dynamic-pricing-outperforms-time-of-use-in-california-ev-charging-pilot-with-98-energy-delivered-off-peak
ev.energy. MCE Sync case study. https://www.ev.energy/case-study/mce-sync
