As electric vehicles (EVs) reach mainstream customers, utilities are facing challenges in serving demand growth from EV charging, particularly on the distribution level. According to recent projections, EV adoption is expected to grow from 4.8 million EVs on U.S. roads today to 78.5 million by 2035 — representing more than 26% of all cars and light trucks. While recent market shifts due to tariffs and potential changes to EV tax credits may affect this projection, utilities are already experiencing substantial impacts directly attributed to EVs, which means that identifying and understanding EV-driving customers has never been more important for utilities to mitigate grid constraints.
In fact, unmanaged EV load has the potential to require billions* of dollars in secondary transformer and service upgrades, all while supply chain constraints on transformers complicate infrastructure expansion. At the same time, additional pressures from building electrification and other load growth are compounding the challenge.
Being proactive is key to reduce or delay the need for such investments, including EV time-of-use rates, active managed charging programs, and other load shifting and shaping interventions. In addition, when machine-learning (ML) and artificial intelligence are applied to AMI data to reveal territory-wide insights, project managers are able to develop strategies to use EVs for load flexibility and virtual power plants (VPPs). AI also empowers utilities to identify areas where EVs are contributing to non-peak time congestion on the distribution grid.
“The rapid growth in EV adoption creates both challenges and opportunities for utilities,” says Brittany Blair, Manager, Research & Industry at the Smart Electric Power Alliance (SEPA) and one of the authors of SEPA’s recent Insight Brief: AI for Transportation Electrification. “Without proactive management, grid impacts from unmanaged charging could be costly, but with the right data-driven approach, utilities can turn EV charging into a grid asset. And the time to prepare for that future is now.”
*California Public Advocates Office (2023); Kevala (2023); NYSERDA (2022)
The Power of AI for Transportation Electrification
U.S. electric utilities have already begun preparing for transportation electrification. As of 2022, 59% of electric utilities in SEPA’s network had established strategic plans for managing this new load, and a key part of those plans is increasing situational awareness of the grid edge and creating data-driven approaches to managing EV charging. What has changed more
recently, is the ability of artificial intelligence (AI) and machine learning to capitalize on existing data flows from the grid and provide the insights needed to enhance utility strategies. At its core, the benefit of using AI for EV detection is precisely that: more sophisticated, data-driven situational awareness.
Artificial intelligence and machine learning provide utilities with an advanced understanding of EV impacts and allow them to better plan for EV demand growth. Using AI, forward-thinking utilities are accelerating their efforts to identify EV-driving customers, create targeted marketing and EV engagement programs, and account for EV charging within their broader distribution system management strategies.
According to SEPA’s research, effectively managing transportation electrification is made easier with a four-step, AI-enabled process:
- Identify Customers with EVs and Their Charging Behaviors: AI enables a more complete understanding of customer EV adoption by identifying EV customers within the utility service territory and parsing data on charging behavior, both of which support utility assessment of system impact and opportunities for managed charging.
- Create Managed EV Charging Programs: With granular information about customer EV adoption and charging patterns, utilities can strengthen programs designed to meet grid needs, avoid grid strain, and deliver potential customer savings.
- Incorporate EVs into Load Forecasts: An AI-based understanding of EV adoption, charging behavior, and grid impact by distribution grid segment (transformer, circuit, substation) allows utilities to better incorporate EVs into load forecasts.
- Map EV Load Growth by Geography: Analyzing EV adoption by geography provides an opportunity to detect local reliability risks that may appear before system-level issues arise, particularly due to high EV penetrations on certain levels of the distribution grid. Early identification can help utilities refine their distribution plans and management strategies.
For transportation electrification specifically, AI solutions that detect EVs from premise-level meter data give utilities visibility into charging behaviors without requiring additional hardware investments. Traditional methods of gathering EV ownership information like vehicle registration records, telematics and customer surveys can leave a data gap that limits utilities’ ability to plan distribution systems efficiently and implement effective EV load management programs. AI can reveal what traditional methods miss.
AI-Powered EV Strategies in Practice
The Insight Brief: AI for Transportation Electrification features two case stories from utilities who are leading the way when it comes to AI-informed EV strategies: Hydro One in Ontario, Canada, and NV Energy in the state of Nevada.
With EV adoption growing, particularly in areas already facing load growth from new construction, Hydro One was seeking better ways to identify EV drivers beyond customer surveys. By implementing Bidgely’s AI-powered analytics, Hydro One was able to identify 20,000 customers with EV charging activity — approximately 10 times more than they had identified via customer surveys. The utility also realized what it calls its “highest click through rate” in recent history via targeted email recruitment campaigns and is able to create territory-specific EV load shapes to inform grid planning.
NV Energy turned to AI to better understand customer preferences and charging trends while identifying and testing technology that could help improve distribution and resource planning processes to prepare for future grid constraints from additional EV load. Using Bidgely’s patented AI, NV Energy was able to detect 50 customers with high-value baseline charging behavior and achieve a load-shift potential of 2-4 kW/vehicle per event – far above the typical 0.2-0.8 kW/vehicle. The utility’s learnings helped inform its 2025-2027 Transportation Electrification Plan.
Looking Ahead
As utilities advance their distribution system planning capabilities for the future grid, improved EV insights can inform a cascade of other investments. Utilities can deploy AI’s capabilities in classification, assessment, automation, prediction, and customer engagement benefitting their teams at every stage from strategic planning to system investments to operations.
“Utilities that invest in managed charging strategies now will have an advantage in navigating the transportation electrification transition before EV adoption reaches a critical point in their territories,” emphasizes Blair. “As EV adoption continues to accelerate, the value of having a utility strategy based on precise, granular charging distribution and load patterns will aid in having a more flexible grid, more targeted infrastructure investments, and overall better customer experience. Data and AI capabilities are one piece of that strategy.”
To read the full utility case stories and learn more about how you can leverage AI to prepare for the growing adoption of electric vehicles, maintain grid reliability and grow customer satisfaction download the full Insight Brief: AI for Transportation Electrification today. And, reach out to our team to schedule a live demo of Bidgely’s EV Solution.
