Carla, a Swedish used-EV sales company, compiled data from 9,954 battery inspections conducted between 2022 and 2026. At the top of the ranking for vehicles with over 100,000 km driven was the Kia e-Niro, maintaining an average State of Health (SoH) of 97.25%. The Hyundai Kona followed at 97.18%, and the Kia EV6 at 95.95%. However, when the same dataset is viewed by mileage segment, the overall average SoH at the 100,000 km mark is 92.4%. This is not a story of "EVs retain over 95% of their capacity even after 100,000 km." What these numbers actually change is not our prediction of battery lifespan, but how we view used EVs.
Heavier Than the Top Score of 97.25%: The Overall Average of 92.4%
Carla listed in its ranking only those models with mileage exceeding 100,000 km and a sufficient number of inspections. The average SoH for the top six models is as follows.
| Rank | Model / Battery | Average SoH beyond 100,000 km |
|---|---|---|
| 1 | Kia e-Niro · 64kWh | 97.25% |
| 2 | Hyundai Kona · 64kWh | 97.18% |
| 3 | Kia EV6 · 77.4kWh | 95.95% |
| 4 | Volvo XC40 Recharge · 69kWh · CATL | 94.70% |
| 5 | Polestar 2 · 78kWh · CATL | 94.35% |
| 6 | BMW i3 · 120Ah | 93.77% |
What this table tells us is that the average was high among the group of top models that met the disclosed criteria. Not all 9,954 inspections covered vehicles with over 100,000 km, and how many units of each model were included is not disclosed either. Even the 20th-ranked Volkswagen ID.3 maintained 91.79%, but this ranking alone cannot yield the average for the market as a whole.
This is where the mileage-segmented data that Carla published earlier becomes useful. Across all 9,954 inspections, the average SoH was 98.6% up to 10,000 km, 94.4% at 50,000 km, and 92.4% at 100,000 km. The decline is not linear. The drop from 70,000 km to 100,000 km was only about 1 point, with the curve flattening out after an initial decline.
However, data becomes sparse beyond that point. At 120,000 km there were only 39 vehicles, and beyond 130,000 km only a handful, so Carla also focused its analysis primarily on the range up to 100,000 km. One cannot extrapolate battery replacement timing by assuming a continued decline at a fixed rate from 92.4%.
Real-World Battery Aging Is More Complex Than Charging Performance
If we try to explain the top three models by saying "lower charging output protected the battery," the EV6 doesn't fit that narrative. According to Kia's materials, the e-Niro's 64kWh battery accepts up to 77kW, taking 54 minutes to reach 80% on a 100kW fast charger. Hyundai similarly states that the 64kWh Kona takes about 54 minutes to reach 80% on a 100kW charger. Meanwhile, the 77.4kWh EV6 supports up to 239kW and can go from 10% to 80% in 18 minutes. Yet it still ranked third with an average SoH of 95.95%.
A vehicle's maximum charging spec does not reflect how many times, or at what power, an owner actually charged it. Thermal management, battery chemistry, vehicle age, and time spent near full charge all affect degradation as well. Since Carla has not disclosed charging history, temperature, or model year by vehicle, no single cause can be isolated from the ranking.
Even in real-world data, charging behavior cannot be ignored. An analysis published by Geotab in January 2026, covering over 22,700 vehicles across 21 models, found an average capacity loss of 2.3% per year. Vehicles heavily reliant on DC fast charging above 100kW showed up to 3.0% loss per year, while those relying mainly on AC or lower-power charging showed about 1.5%, with hotter climates adding a further 0.4 points per year. This is a separate cohort tracked by elapsed years, and not a figure directly comparable to Carla's averages. Even so, it clarifies why the same mileage does not translate to the same SoH.
Laboratory lifespan tests reveal the same complication. A study published in Nature Energy tracked 92 NCA cells intended for commercial EVs over 24 months, and reported that under variable loads resembling real-world driving, lifespan was up to 38% longer than under constant-current testing. Resting cells at high states of charge accelerated cathode capacity loss, and under EV-equivalent loads, calendar aging was also significant. The subject was a single type of cell, and this is not an experiment that explains Carla's ranking. It does, however, demonstrate that mileage, charge cycles, and elapsed time cannot be collapsed into a single metric.
SoH Is Neither Range Nor Failure Rate
SoH represents how much energy capacity remains at the time of inspection, relative to what was available when new. A figure of 97% does not mean that 97% of the rated range always remains. Actual driving range varies with temperature and speed, and is also affected by climate control, tires, and vehicle efficiency. Output performance, cell-to-cell variation, and faults or safety anomalies also cannot be captured by a single capacity ratio.
Carla uses AVILOO equipment to examine five areas: vehicle communication, the battery management system, cell temperature and voltage anomalies, the 12V system, and charging/discharging history. In addition to the SoH displayed on the manufacturer's meter, cell condition and usage history are also part of the evaluation. Carla itself describes the results as "an estimate of condition at the time of inspection."
Disclosure of measurement methodology is essential to interpreting the ranking. AVILOO's FLASH Test records up to about 1,000 signals in roughly 3 minutes while the vehicle is stationary, and is said to have an accuracy of roughly ±3 points compared to the PREMIUM Test, which uses a full discharge cycle. However, Carla does not specify which product was used for each of the 9,954 inspections. Therefore, one cannot apply a ±3 point margin to the model averages, nor can one use a 0.07-point difference to definitively rank the e-Niro above the Kona.
Shifting Used-EV Pricing from Mileage to Individual Inspection
In a brand-level breakdown that Carla added on July 15, the average SoH beyond 100,000 km was 96.8% for Kia, 95.4% for Hyundai, and 95.0% for Mercedes-Benz. Tesla stood at 90.2% and Nissan at 89.3%. But the more data is aggregated at the brand level, the more the age and battery capacity of each manufacturer's various models get averaged together. Differences in thermal management and usage patterns also remain hidden. Carla itself notes that brand-level differences do not guarantee the condition of any specific vehicle, and that differences in model and battery specification need to be considered separately.
This dataset is not a random sample of EVs worldwide. The population consists of vehicles that Carla handled in the Swedish used-car market and that were compatible with AVILOO equipment. The company states that it repairs or replaces batteries that fail inspection, and re-inspects them before sale. Whether the initial failing values or replaced batteries were included in the analysis remains unclear. The cold climate and the particular mix of models in circulation also make generalizing to other regions difficult.
Furthermore, Carla has not published the number of units per model, average vehicle age, mileage distribution, standard deviation, or confidence intervals. An average of 97.25% does not guarantee the condition of every individual e-Niro, nor does it prove that any given e-Niro is always in better condition than a 91.79% ID.3. This is not a peer-reviewed controlled experiment, but a descriptive compilation of inspection records accumulated in the course of used-car sales.
Even so, the 9,954 inspections offer a clear takeaway: judging battery condition by mileage alone at the 100,000 km mark is far too crude. When considering a purchase, buyers should check the SoH certificate alongside battery replacement history, remaining warranty, and fast-charging usage history, and match these against the range and remaining capacity they actually need. Carla has been releasing breakdowns by mileage, model, and brand in stages. If sample sizes and vehicle ages per model, along with how replaced batteries are treated, are also disclosed, this ranking could evolve from promotional material into foundational data for pricing used EVs.