
500,000 battery tests say the pack is fine. The car next to it may not be
Aviloo's used-EV report puts the Ioniq 5 near 94% at 150,000 km and Teslas lower. The 12-point spread inside one model is the number a buyer actually needs.
By EV News Desk
Marcus Berger, who runs Aviloo, the Austrian firm that sells independent EV battery tests to dealers and lessors, looked at more than 500,000 of those tests and refused the comforting headline. "The truth in this data is that averages hide a lot," he said. "Yes, EV batteries are ageing better than the 'they wear out fast' myth suggests — but that's not the most useful thing this report tells you. What matters is that battery health varies significantly between models, and even more between individual cars of the same model."
The first Aviloo Certified EV Battery Report, published 2 September, covers 20 popular models tested between 2022 and 2026 in North and South America, Europe, Australia and Asia. Chinese brands are missing; Aviloo did not have the data. State of health is remaining usable capacity versus new. Medians are given at 50,000 km, 100,000 km and 150,000 km. Sixteen of the 20 models still sat at 90 percent or better at 150,000 km. The other four landed between 85 and 90. Manufacturer warranties typically promise 70 percent at eight years or 160,000 km. The used market, on this evidence, is not full of bricks.
The Ioniq 5 kept more than the Teslas
InsideEVs walked the charts. After 50,000 km the Hyundai Ioniq 5 led, with a median still above 97 percent. The Mercedes-Benz EQA, a compact crossover not sold in the United States, was second at 96.56 percent. Every model in the set cleared 90 percent at that mileage. Strip out two air-cooled outliers — the Renault Zoe and the older Nissan Leaf — and the rest clustered around 94 percent or higher. Aviloo's explanation for the small-pack cars is unromantic: they charge more often to cover the same kilometres, and cycles add up.
At 100,000 km the Ioniq 5 was still first, above 95 percent, with the EQA, Hyundai Kona Electric, BMW i4 and Ford Mustang Mach-E close behind. At 150,000 km — about 93,000 miles — the EQA and Ioniq 5 were virtually tied, 93.97 and 93.79 percent. The i4, Kona and Volvo XC40 Recharge followed. Tesla's two volume cars sat further down. The Model Y median was 90.3 percent; the Model 3 was 88.94 percent. The Mini Cooper SE, Leaf and Zoe were the ones under 88.
That ranking will be quoted in Facebook groups as "Hyundai beats Tesla." It is a median from cars that showed up for an independent test, not a controlled ageing trial. Climate, fast-charging habits and time spent at 100 percent all move the needle. An Aviloo representative told InsideEVs, "Manufacturers also use different battery chemistries with different trade-offs, so even identical charging habits can produce very different results across models." Calendar age does too.
The 12-point gap inside one badge
Berger's sharper point is the spread. For a given model at a given mileage, Aviloo reports the band that covers 99 percent of examples. InsideEVs printed the Model Y window at the 150,000 km mark as 82.9 to 94.9 percent. The i4 ran 87.5 to 96.9. The Mach-E, 85.3 to 96.1. The Model 3, 80.28 to 94.53. Electrive, using the same database, said the Ioniq 5's individual spread at 150,000 km was more than 12 percentage points of state of health, about 42 km of range on that pack.
Two cars with the same badge, similar miles and a similar asking price can therefore be a long-motorway example and a short-hop example. Dashboard health readouts, Aviloo argues, are often a poor substitute for a third-party measurement. That is also how the company makes money, which is worth keeping in the same paragraph as the charts. Across the full 20-model set the medians at 150,000 km land between roughly 87 and 94 percent. The myth of the disposable pack does not survive that spread. The myth of a single number for "the Ioniq 5" does not survive the next one.
What the Guangzhou paper adds, with a caveat
A separate study, published in Scientific Reports in August by Hao Cheng, Zhongwei Gu and Wangqiang Gao at Shanghai Dianji University, tried to isolate driving style from heat, idle time and starting charge. They took a year of 2022 data from 15 EVs in Guangzhou, logged every ten seconds, and sorted behaviour into five styles. Immediate ageing stress in the most aggressive group was 18.4 times the eco group. Average RMS current was 66.02 A versus 20.56 A. Average speeds were similar. The damage was the launches and stops, not cruising fast.
Fed into a degradation model, those profiles projected 21.15 percent capacity loss after 1,000 cycles for the gentlest style and 53.22 percent for the harshest. The Electric Viking, walking through the paper on 2 September, underlined the limit the authors themselves imply: nobody followed these cars for a thousand charge cycles, so those percentages are model output rather than observation. Real packs do not usually fall off a cliff like that model. The direction still matches Aviloo's spread. Two Ioniq 5s can age differently because two drivers exist.
If you are buying used, the practical sequence is dull and sufficient. Prefer a model whose median in this set stayed high. Then ignore the median and test the actual car. Leave the 936 km hypermile videos for the weekend. The battery you are about to finance lives in the 12-point gap Berger was talking about.
More from the file

Sodium-ion batteries are closing in on LFP cost parity, and the last excuse is energy density
CATL expects sodium-ion to match LFP on cost by the end of 2026. The chemistry's remaining problem is density — and two 2025 research papers just attacked it from different angles.

ProLogium's solid-state cell passed the tests. Its factory makes 0.5 GWh a year
A 381 Wh/kg all-solid-state cell is now in commercial production in Taiwan, validated by TUV and UL Solutions. The output is roughly 6,000 car packs a year.

Donut Lab's 409 Wh/kg result is real. It still does not prove the chemistry
VTT measured exceptional energy density from one Donut Battery cell. The report validates the number, but not solid-state chemistry or production readiness.