AVILOO Ranks EQA and IONIQ 5 Top on a Sample Nobody Chose

AVILOO published EV battery health rankings drawn from 500,000 tests that place the Mercedes-Benz EQA and Hyundai IONIQ 5 at the top after 150,000 kilometers and the Nissan Leaf ZE1 at the bottom, but the sample behind the rankings is not random and the underlying percentage figures differ across reports.

A battery diagnostics firm has put a number on one of the most argued-over questions in the electric vehicle market: which packs hold their capacity, and which ones fade. AVILOO's newly published rankings, built from 500,000 tests, place the Mercedes-Benz EQA and the Hyundai IONIQ 5 at the top of the table after 150,000 kilometers of use, and the Nissan Leaf ZE1 at the bottom.

That ordering is the part of the study that travels well. It shows up consistently across independent coverage of the report. The percentage figures behind it do not travel as well — the specific degradation numbers attached to each model vary between accounts of the same study. And the sample itself was not randomly drawn.

Why a Half-Million Tests Is Not the Same as a Half-Million Random Cars

The number 500,000 does heavy rhetorical work. It sounds like the kind of dataset that settles arguments. But sample size and sample selection are separate problems, and only one of them is fixed by volume. A non-random sample that gets ten times larger becomes a more precise measurement of a biased population, not a less biased one.

The mechanism matters here. Battery health tests are not administered to a randomly chosen slice of the fleet. Somebody has to want one. In practice, the people who pay for a third-party diagnostic on an EV pack tend to fall into recognizable groups: sellers preparing a used car for market and hoping for a certificate that supports the asking price; buyers who suspect a problem; owners chasing a warranty claim; leasing and fleet operators running end-of-contract checks; and enthusiasts who simply like data. Each of those motives skews the result in a different direction, and none of them is neutral.

The skew is not necessarily downward, either — which is what makes it hard to correct for. A seller with a strong pack has an incentive to test and publish. A seller with a weak one has an incentive to skip it. That pushes some model cohorts optimistic. Meanwhile, an owner who suspects degradation has a strong reason to test, which pushes other cohorts pessimistic. Whether a given model looks good or bad depends partly on which of those behaviors dominates among its owners, and that varies by vehicle age, resale market, warranty terms and country.

What the Rankings Can and Cannot Support

None of this makes the study worthless. It makes it a different kind of evidence than a controlled fleet trial. Read carefully, the AVILOO data supports statements about the population of tested vehicles: among cars whose owners chose to have the pack measured, EQAs and IONIQ 5s at high mileage came out best and Leaf ZE1s came out worst. That is a real finding about a real dataset.

What it cannot support without further work is the stronger claim that gets made on its behalf: that an EQA bought today will outlast an equivalent Leaf by a specific margin. Getting from the first statement to the second requires knowing that testing propensity is roughly equal across models, and there is no obvious reason to assume that. As Energy News (H2) notes, the sample is not random, and the percentage figures circulating alongside the ranking are inconsistent from one report to the next.

That inconsistency is its own warning. When the ordinal result — first, second, last — is stable across coverage but the cardinal result is not, it usually means the ranking is robust to methodology and the numbers are not. Readers should treat the podium as the finding and the decimals as provisional until AVILOO publishes the per-model sample counts, the mileage distribution within each cohort and the test protocol behind each figure.

The Leaf's Position Has a Known Engineering Explanation

The bottom placement of the Nissan Leaf ZE1 is the least surprising line in the table, and that is worth saying plainly rather than treating it as a scandal. Early Leaf generations are the most-discussed case study in EV thermal management precisely because they lacked active liquid cooling of the pack, and heat is the primary accelerant of lithium-ion capacity loss. A ranking that put the Leaf ZE1 at the top would have raised more methodological questions than one that puts it at the bottom.

Conversely, the models at the top of the list are newer platform designs. Comparing them to an older architecture at the same 150,000-kilometer mark compares two things at once: chemistry and cooling, but also calendar age, charging habits of the era and the software that manages state of charge. Degradation is a function of time as well as distance, and a table organized by kilometers alone folds those variables together.

Why Residual Values, Not Bragging Rights, Are the Real Stake

Battery health data is drifting from enthusiast curiosity toward market infrastructure. Used EV pricing, lease residuals, insurance write-off decisions, warranty reserves and the economics of second-life stationary storage all depend on a credible answer to how much capacity a pack retains at a given age. Whoever supplies that answer at scale acquires real influence over how much used electric cars are worth — and over which brands carry a degradation discount.

That is the reason to be strict about sampling rather than dismissive of it. A ranking that is directionally right and numerically shaky can still move prices, because dealers and lenders need a number and will use the one that is available. The corrective is disclosure: cohort sizes, selection criteria, geographic spread and how repeat tests on the same vehicle are handled.

Automakers, for their part, hold the cleanest data in the industry — full telematics histories on every car they have sold — and disclose comparatively little of it. Until that changes, third-party datasets like this one will keep filling the gap, with all the selection problems that come from measuring only the cars whose owners asked to be measured.

The Wider Tape

The report lands in an unremarkable market week rather than a dramatic one. At the last close on Friday, September 4, 2026, the S&P 500 tracker (NYSEARCA: SPY) finished at $770.23, down 0.38% from the prior close of $773.17, while the Nasdaq 100 fund (NASDAQ: QQQ) ended at $719.00, up 0.19%. The Dow 30 vehicle (NYSEARCA: DIA) closed at $534.12, off 0.52%. Battery-durability findings rarely register on an index in a single session; they show up later, in residual value tables and warranty accruals.

What to watch next: whether AVILOO releases per-model sample counts, whether the conflicting percentage figures get reconciled to a single published set, and whether any automaker responds with telematics data of its own. The first would let the ranking be checked. The last would make it unnecessary.

Key facts

• Tests in dataset: 500,000
• Mileage benchmark: 150,000 kilometers
• Top-ranked models: Mercedes-Benz EQA and Hyundai IONIQ 5
• Benchmark close (Sept 4, 2026, 20:00 GMT): SPY $770.23, -0.38%

Frequently asked questions

What did AVILOO's study find?

AVILOO published EV battery health rankings drawn from 500,000 tests. The Mercedes-Benz EQA and Hyundai IONIQ 5 placed at the top of the table after 150,000 kilometers of use, while the Nissan Leaf ZE1 placed at the bottom. That ordering appears consistently across independent coverage of the report.

Why does it matter that the sample is not random?

Because the people who pay for a battery test choose to do so for a reason — selling a car, suspecting a fault, chasing a warranty claim. Those motives differ by model and market, so the tested cars may not represent the wider fleet. A larger non-random sample measures a biased population more precisely rather than removing the bias.

Are the percentage figures in the ranking reliable?

Less so than the ranking itself. The specific degradation percentages attached to each model do not hold up consistently across reports of the same study, even though the order of finishers does. That pattern suggests the ordinal result is robust to methodology while the exact numbers should be treated as provisional.

Why does the Nissan Leaf ZE1 rank last?

Early Leaf generations are the standard case study in EV thermal management because they lacked active liquid cooling of the battery pack, and heat is the main accelerant of lithium-ion capacity loss. A bottom placement is consistent with that known engineering difference rather than an anomaly in the data.

Does the ranking prove one brand's batteries are better?

Not on its own. Drawing that conclusion requires assuming that owners of every model are equally likely to have their packs tested, which is unlikely. The study supports statements about the population of tested vehicles; extending it to all cars of a given model needs per-model sample counts and selection details AVILOO has not detailed publicly.

Why is battery health data commercially important?

Used EV prices, lease residual values, insurance write-off decisions, warranty reserves and second-life stationary storage economics all depend on how much capacity a pack retains at a given age. Whoever supplies that figure at scale influences what used electric cars are worth and which brands carry a degradation discount.

Sources

• AVILOO’s 500,000-Test EV Battery Study Comes From a Sample That Isn’t Random — Energy News (H2)

Photo: Sergey Meshkov · Pexels Licence — source