The short answer: the average electric car loses about 2.3% of its battery capacity a year, which puts it near 82% after eight years — still comfortably usable. The rate is not fixed: cars doing most of their charging on high-power DC degrade at up to 3.0% a year, roughly double those charged mainly at home on AC. How you charge matters more than which car you bought.
What the fleet data actually says
Two datasets are worth taking seriously, because both are built on large numbers of real vehicles rather than laboratory cycling or manufacturer claims.
Geotab: 22,700 vehicles, 21 models
Geotab’s 2026 analysis puts average annual degradation at 2.3%, projecting to about 81.6% of original capacity after eight years. The number worth pausing on is that this is up from the 1.8% they reported in 2024 — not because batteries got worse, but because usage changed. Geotab attributes the increase largely to growing reliance on high-power DC fast charging.
Their split by vehicle class is also useful: multi-purpose vehicles and SUVs average about 2.7% a year, lighter cars about 2.0%. Those are the two figures this site uses on individual vehicle pages, labelled as class averages rather than measurements of a specific model.
Recurrent: over a billion miles of driving
Recurrent’s 2026 market report, drawn from more than a billion miles of real-world driving, found the average EV retains about 97% of expected range after three years and 95% after five. Sixty-eight percent of model-year 2023 cars still meet or exceed their original EPA range estimate today.
The figure that most changes the conversation, though, is the replacement rate: 0.3% among 2022-and-later vehicles. Three cars in a thousand. Battery replacement is a real risk to price into a purchase, but it is a tail risk, not a routine maintenance item.
What that means in practice
| Years owned | Car (2.0%/yr) | SUV or MPV (2.7%/yr) | What it feels like |
|---|---|---|---|
| 3 | 94% | 92% | Barely detectable. Within the noise of weather and driving style. |
| 5 | 90% | 86.5% | Noticeable on a long trip. One more charging stop on a 700 km drive. |
| 8 | 84% | 78.4% | Real. A 450 km car is now a 350–380 km car. Still above every warranty floor. |
| 10 | 80% | 73% | The point at which a long-distance driver starts thinking about the next car. |
Straight-line projection from the class averages above, as used on our vehicle pages. Real degradation is faster in the first months and then flattens, so these figures are slightly pessimistic in the middle years.
What actually drives it
1. How much of your charging is high-power DC
This is the largest controllable factor by a wide margin. Vehicles using DC above 100 kW for more than about 12% of their charging sessions degraded at up to 3.0% a year. Vehicles charged primarily on AC came in nearer 1.5%. Over eight years that is the difference between roughly 76% and 88% of original capacity — on the same car.
This is also why the honest answer to “which car degrades least” is usually “the one you can charge at home”.
2. Thermal management
Every car in our comparison tool uses active liquid cooling, which is now effectively standard. Where it is absent — older or cheaper designs using passive or forced air — degradation is dramatically worse in hot climates, and the difference is large enough that we weight it as the single heaviest input in our battery safety score.
3. Time, independently of mileage
Calendar ageing happens whether you drive or not, which is why published fleet studies report degradation per year rather than per kilometre, and why this site does the same. A three-year-old car with 20,000 km has lost more capacity than a naive per-kilometre model would predict.
4. State of charge at rest
Nickel-rich chemistries age faster held near 100%. LFP does not care, which is one of the practical arguments for it — see our chemistry comparison.
What nobody publishes, and why we do not either
You will find sites listing degradation model by model. Treat them with suspicion. Geotab anonymises the models in its published findings. No manufacturer releases fleet degradation data. No independent body measures it at the scale required to separate one model from another with statistical confidence.
So on this site, two SUVs show the same annual figure — because we genuinely do not know which of them ages faster, and inventing a difference would be worse than admitting the gap. Every degradation figure here is labelled as a class average, with a link to its source.
Warranties: what they actually promise
Eight years or 100,000 miles with a 70% capacity floor is the industry norm, and in the United States it is also the legal minimum. Hyundai and Kia go to ten years, the longest among high-volume brands.
Read the floor carefully. A warranty that replaces the pack below 70% is not a promise that you will keep 70%; it is a promise about the point at which something has clearly gone wrong. At 2.3% a year an average car reaches 82% at eight years — comfortably outside warranty territory, which is exactly why claims are rare.
Common mistakes
- Reading a cold-weather range drop as degradation. A car showing 30% less range in January has not lost 30% of its battery. Cold reduces available energy temporarily and it comes back.
- Trusting the dashboard range estimate. It is a prediction based on recent driving, not a measurement of battery health. Judge degradation from energy consumed over a full charge, or a proper diagnostic.
- Extrapolating from the first year. Degradation is fastest in the early months and then flattens sharply. A 4% first-year loss does not mean 40% at ten years.
- Assuming replacement means a whole new pack. Module-level repair is increasingly normal and costs a fraction of a full replacement.
- Buying a bigger battery “to be safe”. A bigger pack costs more to buy, insure and replace, and does not degrade more slowly. Buying an efficient car does more for long-term range than buying a large one.
Practical guidance
- Charge on AC at home where you can. This is the single biggest lever, and it is worth more than any specification difference between cars.
- Use DC fast charging when it saves you time, not by default. Occasional rapid charging on trips is fine. Doing it several times a week because it is convenient is what the Geotab data is picking up.
- On a nickel-chemistry car, set a daily limit around 80% and charge fully only before a long drive. On LFP, charge to 100% and stop worrying.
- Do not leave the car at 100% or near 0% for weeks. Around 50% is the kindest state for long storage.
- Buying used? Ask for a battery health report rather than reading the range estimate. And weight a car’s charging history — a former fleet or rideshare car has probably lived on DC.
Frequently asked questions
How long will an EV battery actually last?
At current fleet-average rates, a battery is at roughly 82% after eight years and around 80% after ten. Cars are being retired for other reasons long before the pack becomes unusable, and the 0.3% replacement rate among recent model years reflects that.
Does fast charging really damage the battery?
Heavy reliance on it does. The measured gap is roughly 3.0% a year for cars doing more than about 12% of sessions on DC above 100 kW, against roughly 1.5% for AC-primary vehicles. Occasional trip charging is not the problem; making rapid charging your default is.
What does replacing a battery cost?
Out-of-warranty replacement was quoted at roughly $130–$200 per kWh installed in 2026, so a 75 kWh pack lands around $10,000–$15,000. Remanufactured packs undercut that by thirty to fifty percent, and module-level repair costs far less again. Each vehicle page here shows an estimate with the formula published.
Is degradation getting better or worse?
The measured average went up between 2024 and 2026 — from 1.8% to 2.3% a year. Cells did not get worse; charging behaviour changed as fast-charging networks got denser and drivers used them more.
Should degradation change which car I buy?
Less than you would think. Choose on efficiency, charging speed, thermal management and whether you can charge at home. Those decide both your running costs and your degradation rate. Our running cost calculator shows projected battery health at the end of your ownership alongside the money.
Chemistry sets the ceiling on how well a pack ages, and the two families behave differently enough to be worth reading separately: LFP cycles far more times but suffers more in the cold, while NMC and NCA trade some of that longevity for energy density.
Sources
- Geotab, EV battery health: findings from 22,700 vehicles (2026) — annual degradation rate, class split, DC fast-charging effect
- Recurrent, 2026 EV Market & Trends Report — range retention at three and five years, battery replacement rate
- 2026 manufacturer battery warranty comparison — terms, mileage caps and capacity floors
- 2026 out-of-warranty replacement cost surveys — per-kWh installed pricing