When an odometer reads 150,000 kilometers, how worn out is the heart of that vehicle? Drivers who have cycled through gasoline-powered cars for years learned to gauge overall wear and tear from mileage alone, factoring in metal component fatigue and oil degradation. But in the world of electric vehicles (EVs), which carry a massive lithium-ion battery pack beneath the floor, that intuition no longer applies.
The fear that the single most expensive core component—the battery—might become unusable after just a few years has long discouraged new-car purchases and distorted pricing in the used-car market. Most automakers guarantee free battery replacement if capacity falls below 70% of the original within 8 years or 160,000 kilometers (100,000 miles). That 70% threshold, however, has also reinforced consumer suspicion that a meaningful number of vehicles might degrade rapidly enough to approach that boundary.
A large-scale study published on September 2, 2026, by Aviloo, an Austria-based EV battery diagnostics company, offers numbers that challenge this pessimistic view. Titled the "AVILOO Certified EV Battery Report," the study draws on used-EV battery test data the company accumulated between 2022 and 2026. Analyzing more than 500,000 readings across 20 representative models popular in Europe and elsewhere, the report found that the median battery health at the 150,000-kilometer mark fell between 87% and 94% for most models.
Still, these statistics demand a skeptical eye. This is not a peer-reviewed academic paper—it is a commercial report published by a private company whose core business is paid battery diagnostics and certification, framed in a way that promotes the value of its own diagnostic services. Before taking comfort in the reported capacity retention, we need to scrutinize both the limitations of the measurement methodology and the troubling degree of "individual variation" hidden behind these figures.
- Lower bound across 20 models
- Upper bound across 20 models
データを表で見る
| Lower bound across 20 models (%) | Upper bound across 20 models (%) | |
|---|---|---|
| At 50,000 km | 91.2 | 97.1 |
| At 100,000 km | 88.3 | 95.3 |
| At 150,000 km | 86.7 | 93.8 |
As this chart shows, when looking only at overall medians, the rate of capacity decline as mileage increases is remarkably gentle. But this smooth curve conceals a fundamental pitfall for anyone shopping for a used EV.
How a Three-Minute Data Pull Estimates Battery Lifespan
To rigorously understand what is happening inside a battery that has been driven for years, one would ideally need to fully discharge cells in a laboratory setting and measure the integrated current and heat behavior over long charge-discharge cycles. But such physical testing is impractical at a neighborhood used-car dealership or repair shop.
Aviloo's diagnostic method, called the "FLASH Test," involves plugging a dedicated device into the onboard diagnostics (OBD) port near the driver's seat and pulling internal data from the vehicle's electronic control unit (ECU) in about three minutes. During this brief communication window, the tool retrieves signals held by the battery management system (BMS)—cell-to-cell voltage variance, charge-discharge cycle counts, cumulative energy counters, and historical temperature data.
The resulting State of Health (hereafter SoH) figure is not simply a transcription of the number displayed by the vehicle's own BMS. Instead, Aviloo cross-references the data against its global database of more than one million cumulative tests, comparing vehicles with similar mileage, age, and usage conditions, and applies a proprietary statistical algorithm to derive an estimate. The company defines SoH as "the ratio of currently available battery capacity to the effective capacity available when the vehicle was new." An SoH of 95%, it says, can be interpreted as meaning the vehicle retains roughly 95% of its original single-charge range.
This measurement approach carries inherent uncertainty. For each model, Aviloo calculates not just a median but a prediction interval intended to encompass 99% of vehicles (a quantile curve with a confidence coefficient alpha of 0.99). However, vehicles sharing the same model name can differ in battery capacity or cell supplier depending on model year. According to the company's public materials, multiple battery-capacity variants under a single model name are consolidated into one category for analysis—meaning some of the variance seen in the report stems not purely from degradation differences but from underlying specification differences.
Most importantly, the algorithm itself—applied to a dataset spanning North America, South America, Europe, Australia, and Asia—remains proprietary and undisclosed. No independent academic institution has conducted an objective replication or validation of the algorithm. Furthermore, data from Chinese-made EVs—despite China being the world's largest EV market with many distinct battery chemistries—was excluded entirely from this analysis due to an insufficient sample size. The figures presented should be evaluated strictly as an aggregate of commercial diagnostic data collected in specific market environments, primarily in Europe and North America.
The 13.5-Point Gap That Undercuts the Comfort of the Median
Looking at the four representative models for which Aviloo provided detailed figures reveals both the pace at which batteries degrade with mileage and the striking degree of divergence between individual vehicles of the same model.
データを表で見る
| Median SoH at 150,000 km (%) | |
|---|---|
| Tesla Model Y | 90.1 |
| VW ID.4 | 90.9 |
| Hyundai Ioniq 5 | 93.4 |
| Nissan Leaf ZE1 | 85.9 |
The Tesla Model Y (78.8 kWh) shows a solid trajectory: a median SoH of 94.5% at 50,000 km (roughly 379 km of usable range), 91.8% at 100,000 km, and 90.1% at 150,000 km (about 361 km of range). The Volkswagen ID.4 (77 kWh) follows a similar path, declining from 95.4% at 50,000 km to 90.9% at 150,000 km (usable range falling from roughly 373 km to 355 km). The Hyundai Ioniq 5 (72.6 kWh) maintains even higher figures—according to the company's own blog post, 97.3% at 50,000 km and 93.4% at 150,000 km (range dropping from roughly 335 km to 321 km). Notably, a separate press release from the same company lists slightly different figures—97.2% at 50,000 km and 92.8% at 150,000 km—showing a minor discrepancy even within the automaker's own published materials.
In contrast to these vehicles with large battery packs, the Nissan Leaf ZE1 (40 kWh)—one of the pioneers of the EV market—posted the weakest results. Vehicles with smaller battery capacity must cycle through more charge-discharge cycles to cover the same distance. The Leaf ZE1's median SoH fell from 90.8% at 50,000 km to 87.6% at 100,000 km and down to 85.9% at 150,000 km. Its usable range dropped from about 195 km at 50,000 km to about 185 km at 150,000 km.
| Model (representative battery capacity) | Median SoH at 50,000 km | Median SoH at 100,000 km | Median SoH at 150,000 km | Change in usable range (approx.) | Spread within same model at 150,000 km |
|---|---|---|---|---|---|
| Tesla Model Y (78.8 kWh) | 94.5% | 91.8% | 90.1% | ~379 km → ~361 km | Up to 11.0 points |
| Volkswagen ID.4 (77 kWh) | 95.4% | 92.7% | 90.9% | ~373 km → ~355 km | 11.0–12.0 points |
| Hyundai Ioniq 5 (72.6 kWh) | 97.3% | 95.1% | 93.4% | ~335 km → ~321 km | Over 12.0 points |
| Nissan Leaf ZE1 (40 kWh) | 90.8% | 87.6% | 85.9% | ~195 km → ~185 km | Up to 13.5 points |
The real takeaway from this comparison is not the height of the median values themselves. What deserves the most attention is the widening "spread within the same model," shown in the rightmost column.
At 50,000 km, the SoH spread within a given model ranged from 7.4 to 10.0 percentage points. But by the time vehicles reach 150,000 km, that gap widens sharply. The Leaf ZE1, which recorded the largest divergence among all 20 models studied, showed up to a 13.5-point difference between its best-preserved and most-degraded individual vehicles—equivalent to a practical difference of roughly 29 km in range per charge. According to Aviloo's broader analysis, when accounting for the 99% interval across the population, the gap can widen to as much as 16.4 points in some cases.
For the ID.4 as well, the range difference between the best- and worst-performing vehicles at 150,000 km reached up to 44 km. The Model Y showed variation of up to 11 points, and the Ioniq 5 exceeded 12 points.
Two used cars assembled at the same factory, fitted with batteries from the same supplier, and driven the identical 150,000 kilometers can end up in vastly different conditions: one retaining nearly 95% of its original performance, the other worn down into the low 80s. Simply glancing at the odometer offers no way to know where any given vehicle falls within that spread.
Where Observed Correlation Ends and Proven Mechanism Begins
Why does such a wide gap emerge between vehicles that have covered the same distance? Marcus Berger, CEO of Aviloo, offered a clear perspective in interviews with various media outlets. "With an internal combustion car, you could roughly gauge a vehicle's value and mechanical condition just by looking at its model year and mileage. That's not the case with EVs. Two cars of the same year and mileage can look identical, yet you have no way of knowing how each one was actually treated."
Berger identified two primary external factors that accelerate degradation: climate conditions and how the battery's charge level is managed while parked. The company's fleet data revealed a statistical trend showing that vehicles used in regions with higher average annual temperatures tend to degrade faster. Berger also addressed everyday parking habits, warning, "You should never leave a car sitting at a high charge level—100%, 90%, or even 80%. That alone damages the battery," and recommended keeping the state of charge (hereafter SoC) between 30% and 70% during long periods of parking.
Here, scientific rigor demands a clear distinction: what the company presented is an "observed correlation" drawn from real-world big data, not a specific degradation mechanism isolated and proven in a controlled laboratory setting.
It is well established in lithium-ion battery electrochemistry research that high SoC combined with high ambient temperature accelerates metal-ion dissolution from the cathode active material, oxidative decomposition of the electrolyte, and film formation on the anode surface. Berger's comments align with this established physical chemistry, but Aviloo's own 500,000-record report did not involve a controlled comparative experiment that strictly isolated variables such as driving patterns, fast-charging frequency, or ambient temperature profiles.
The Leaf ZE1's notably faster degradation likely stems not only from the higher number of cycles required due to its smaller battery capacity, but also from the structural difference between its air-cooled thermal management system and the liquid-cooled active thermal management systems adopted by most recent EVs. However, the report does not quantitatively isolate how much this thermal-design difference contributed to the observed statistics. While the executive's advice carries value as a rule of thumb drawn from experience, it should be understood as distinct from a scientifically proven causal relationship.
Market Opacity and the Business Interests Behind the Push for Inspection
The reality this study presents—a median of 87% to 94% capacity retention—invites a reassessment of existing warranty frameworks. The common industry standard of guaranteeing 70% capacity retention over 8 years or 160,000 kilometers appears, based on this study's findings, to function as a conservative safety net with considerable margin for most users. Barring extreme abuse or early manufacturing defects, the statistical probability of falling below that 70% warranty threshold through normal driving is quite low.
However, there remains a deep gap between simply staying within warranty terms and being recognized as having fair market value. Current vehicle appraisal systems have been built over decades around mileage and initial registration date for internal combustion vehicles. The fact that internal battery degradation cannot be inferred from a car's exterior or odometer reading creates real psychological burden for both used-car dealers and everyday consumers—like carrying an invisible risk.
To address this opacity, calls for a standardized battery health certification are rapidly gaining momentum within Europe's auto industry. Three major UK-based EV industry organizations have petitioned the government to establish mandatory battery health certification requirements for used EV transactions. James Strong, manager of Tesla UK's certified pre-owned sales division, has similarly argued that establishing a nationally recognized battery SoH standard would dramatically boost consumer confidence and transparency in the used EV market.
Beneath the push for stronger regulation and standardization lie clear commercial interests held by the companies providing diagnostic technology. In the summer of 2026, Aviloo launched a "battery warranty" business across Europe, offering monetary guarantees to used EV buyers based on its own diagnostic results. The company has also announced plans to publish its "Certified EV Battery Report" twice yearly going forward.
Aviloo's FLASH Test diagnostic technology has been adopted or recognized by influential third-party automotive organizations such as Germany's ADAC and Austria's ÖAMTC, and its practical standing within the industry is becoming well-established. That said, the "independence" the company claims refers only to its neutrality from any specific automaker—not to validation by an unaffiliated academic institution conducted free of charge. Readers should always keep in mind that this data comes from a company with a direct financial incentive to boost the market value of its own diagnostic and certification services.
The statistical fact that roughly 90% of battery capacity survives even after 150,000 kilometers on the odometer is, without question, a compelling testament to how far the technology has come. But behind that reassuring average lies the sobering reality of individual variation reaching as much as 13.5 percentage points. Was the vehicle frequently fast-charged? Was it left sitting fully charged under the blazing summer sun for weeks at a time? Or was it kept in a mild climate with careful charge management? Until a unified verification standard emerges capable of decoding this black box of past usage history, the lingering uncertainty felt by used EV buyers is unlikely to fully disappear.