A used EV inspection should treat the high-voltage battery as a measured system, not a mystery box. The dashboard range estimate alone is weak evidence because it changes with recent speed, climate control, temperature and driving style. Better evidence combines state-of-health data where the manufacturer exposes it, a full-charge usable-energy or range test, charging behavior, diagnostic codes, warranty status, and the car’s climate and fast-charging history.
Begin with the exact battery and warranty, not the model name
Two cars with the same badge can have different pack sizes, chemistries, software revisions or warranty terms. Decode the VIN and trim, then pull the manufacturer’s warranty document for that model year. The U.S. Department of Energy notes that several manufacturers offer roughly 8-year/100,000-mile battery warranties, but terms vary and some capacity guarantees use a specific threshold. Confirm whether the warranty transfers, its in-service start date, mileage limit, exclusions and the process used to measure capacity.
State of health is useful only if you know how it was produced
A reported 91% state of health sounds precise, but ask what system calculated it. Some manufacturers expose a service diagnostic; some apps estimate from battery-management data; some owners divide displayed usable energy by an assumed new capacity. Record the tool, software version, temperature and state of charge. Treat an app estimate as one data point unless the manufacturer accepts that measurement for warranty decisions.
A third-party connected-data report can add context, especially when it compares the car with the same model and age, but it is not the same thing as the manufacturer’s warranty test. Recurrent, a used-EV battery-data specialist, explicitly says its range estimates are statistical and do not guarantee battery health. Use a report like that as another independent signal; when warranty eligibility or a large price adjustment depends on the result, ask for the manufacturer-accepted diagnostic procedure.
| Check | Good evidence | Weak evidence |
|---|---|---|
| Capacity | Manufacturer diagnostic or repeatable measured usable kWh | Guess from dashboard range alone |
| Charging | Observed AC/DC charging session at expected power after normal taper behavior | Seller says “charges fast” |
| Thermal system | No cooling faults, normal temperatures, service history | No warning light during a five-minute drive |
| Warranty | Model-year warranty document + in-service date + mileage | “All EV batteries are 8/100” |
| Range | Controlled route with known starting/ending SOC and weather | One optimistic range estimate after reset |
Run a road test that gives the battery enough work to reveal itself
Start at a known state of charge, photograph the odometer and energy screen, reset the trip meter, then drive a mixed route long enough to use a meaningful portion of the pack. Record distance, energy consumption, temperature and ending state of charge. You are not trying to create a laboratory capacity test on public roads. You are looking for coherent behavior: stable power, no sudden state-of-charge jumps, no battery or cooling warnings, and efficiency that makes sense for the weather and speed.
Test charging because a healthy-looking pack can still have charging faults
If practical, connect to both Level 2 AC and a compatible DC fast charger. Confirm the port locks, session starts normally, and the car accepts power consistent with its charge curve and current state of charge. Fast-charge power naturally tapers as the battery fills, so a low number at 85% is not the same as a fault at 20%. Check for error messages, damaged pins, heat damage or adapters of uncertain quality.
Climate history changes what “normal degradation” looks like
DOE’s Alternative Fuels Data Center explains that battery life is affected by climate, charging and driving patterns, chemistry, design and thermal management. A car that spent years in extreme heat may age differently from the same model in a moderate climate. Conversely, temporary winter range loss does not equal permanent capacity loss. Look for where the car lived, how it was stored, and whether its model uses active pack cooling or heating.
Failure and degradation are also different risks. EPA’s 2026 EV-myths review cites a study of roughly 15,000 vehicles in which battery replacements due to failure averaged 2.5% across early models through model year 2023, excluding major recalls, and were below 0.5% for model years 2016 onward. That does not tell you the usable capacity of this car; it is a reminder not to treat normal range loss as proof that the pack is about to fail.
Replacement cost is model-specific—do not shop from scary generic numbers
Battery replacement can be expensive, but the correct number depends on whether a module can be repaired, whether a remanufactured pack exists, labor, pack size, warranty and manufacturer parts pricing. Obtain a current quote for the exact model if the pack is near a warranty threshold or diagnostic data is poor. A low purchase price is not automatically a bargain if the battery has little remaining usable energy and no economical repair path.
The rest of the car still matters
EVs can have expensive tires, suspension wear, collision damage, cabin electronics, heat pumps, onboard chargers and driver-assistance hardware. Regenerative braking can reduce friction-brake use, which can also allow rotors to rust if seldom exercised. Inspect underbody shielding for impact damage near the pack. Check recalls by VIN through NHTSA and confirm whether any battery campaign or software update is open.
A strong used-EV purchase therefore has several independent pieces of evidence pointing the same way: plausible battery health, predictable range, normal charging, no thermal or high-voltage faults, documented warranty status, and a conventional mechanical/body inspection. If the seller will show only a 100% dashboard range number and refuses diagnostics or charging tests, the uncertainty itself belongs in the price—or in your decision to walk away.