Public Infrastructure in the Dark: The Hidden Reliability Challenge Behind America’s EV Charging Network
By Jonathan Shan, Summer 2026 CGS Student Research Assistant, Poolesville High School
Public EV charging infrastructure is expanding rapidly across the U.S., but much less is known about what happens to charging stations after they are installed. Working with Jiehong Lou, Associate Research Professor in the School of Public Policy and Assistant Director at the Center for Global Sustainability, I analyzed the lifecycle of public EV charging stations, from when they first appear in the network to when they eventually disappear from the data. Using three years of near-real-time status data covering more than 82,000 charging station IDs nationwide, we examined what happens in the months before a station drops out of the tracked network. One of the clearest patterns was unexpected: usage stayed relatively stable. Drivers didn't gradually use these stations less before they disappeared. What changed was the station’s ability to report its status.
During the twelve months before a station disappeared, the share of time it failed to report basic status information rose from about 23% to 73%. Many stations seemed to start going quiet in the data long before they vanished from the record. That matters because much of the conversation around EV infrastructure focuses on building more chargers and keeping them available to drivers. Our preliminary results suggest that another problem may be developing behind the scenes. Software, network connectivity, monitoring systems, or other operational issues may begin breaking down while drivers continue to use these chargers.
The differences between states were also large. Among stations installed early enough to follow over time, roughly a quarter eventually disappeared from tracking nationwide. Some of that variation may reflect when different states began expanding their charging networks, which requires additional investigation.
We also examined whether station disappearance was associated with local income, race and ethnicity, education, EV adoption, station size and type, and previous performance. One of the most important questions from this work is whether enough attention is given to the integrity of the digital infrastructure behind the physical infrastructure for EV charging stations. That information matters not only for maintenance, but also for understanding whether investments in charging infrastructure continue delivering value over time.
If reporting failures consistently appear before stations disappear, that missing data may provide an early warning. Instead of noticing a problem only after a charger drops out of the network, operators and public agencies could use changes in reporting behavior to identify stations that need attention earlier. We used a machine-learning model to predict which stations may be at risk of disappearing based on changes in reporting behavior, previous performance, and other station characteristics. If operators and policymakers can detect those warning signs while a charger is still active, they may have time to investigate before the station drops out of the network.
As the United States continues expanding its charging network, building chargers is only the first step. We also need to understand whether they remain visible, reliable, and usable after installation.
