Why Your Parked EV's Battery Percentage Can Rise

You park your EV at 58%. Hours later, with no charger ever connected, the number is higher. That is not the battery inventing energy — it is the dashboard's estimate catching up with the battery's real state. Battery percentage is not a fuel-tank reading; it's an estimate the battery management system (BMS) builds from current, terminal voltage, temperature and a model of how the cell behaves. While current is flowing, the loaded terminal voltage isn't the same thing as the battery's equilibrium open-circuit voltage. Stop the current and that load-dependent voltage drop disappears almost immediately — but inside the electrode, lithium concentration gradients built up during discharge don't vanish with it. They keep redistributing by diffusion, the surface lithiation keeps changing, and the electrode's equilibrium potential keeps evolving — a process that peer-reviewed measurements show can continue for minutes to hours, depending on chemistry, state of charge, temperature and history. Stopping current is not the same thing as reaching electrochemical equilibrium. That relaxing voltage gives the BMS's running estimate a new piece of evidence — not a reset, not a direct voltage-to-percentage lookup, but another constraint the estimator weighs against its existing model. If that evidence disagrees with the running estimate, the BMS may revise the displayed number, and that revision can go up or down depending on the direction and size of the disagreement. There's a concrete real-world instance of this kind of rest-based check: one Tesla Model 3 service procedure tells technicians to wait 20 minutes after discharging the pack and re-verify state of charge, because open-circuit-voltage settling matters to an accurate reading — though that 20-minute figure is a service-shop procedure, not a universal timer for how long a parked car's own relaxation takes. Two more variables shape how visible all this is. LFP cells have a flatter voltage-vs-SOC curve over much of their range, so the same small voltage uncertainty can translate into a bigger SOC uncertainty — that doesn't mean LFP cars jump more often while parked, just that voltage is a blunter instrument for them. And temperature changes the battery's relaxation behavior, which changes the measurement conditions the estimator has to interpret — without dictating which way the number moves. Separately, a parked EV can also spend real energy: onboard electronics, low-voltage support and thermal management can draw from the traction battery even with no charger plugged in, so "no charger connected" never means "no energy moved" in either direction. Put together: the car stopped, but the battery didn't instantly reach equilibrium. Its internal state kept relaxing, the voltage kept moving, and the BMS could revise what it had inferred. Current: zero. Voltage: still moving. The percentage is a view of the battery, not the battery itself. Sources: Theuerkauf & Swan, Batteries 8(8):77 (2022) DOI 10.3390/batteries8080077 - Bharathraj, Adiga & Mayya, iScience 28:113769 (2025) - Jorkesh, Akbari, Ahmed & Habibi, Journal of Power Sources 682:240338 (2026) - Tesla Model 3 Service Manual, HV Battery Discharge section - Tesla Motors US Patent 8004243B2 - Nature Communications, nature.com/articles/s41467-025-56485-7 - Tesla Model 3 Owner's Manual - Polestar 2 Manual (2026). Full citations, URLs and evidence classes are recorded in this episode's production source ledger and claim ledger. Video credits: Car parked in garage - Baran Robin, Pexels License. Tesla interior - Pexels License. Tesla charger plug-in - mygiguser, Pexels License. Gear-pump cavitation - Sir Balefire, CC0, Wikimedia Commons. Hydraulic aggregate - Barbara Rezar Grilc, CC BY 3.0, Wikimedia Commons (attribution required). This channel is not affiliated with, endorsed by, or sponsored by Tesla, Polestar, or any other organization named in this video. Corrections policy: if we get something wrong, we pin the correction.