Cars & Used-Car Buying

Buying a Used EV? Ask What the Battery Report Actually Measures

By Priya Lindqvist / October 20, 2026 / Updated November 8, 2026

White electric car connected to an outdoor charging station

Two used EVs with the same odometer reading can carry very different batteries, and the dashboard range figure will not settle which is which. Sellers and dealers increasingly hand over a battery report to fill that gap. The catch: “battery health” can mean a direct test, an estimate built from charging data, or a statistical projection from similar cars. They are not interchangeable, and a bare percentage doesn’t tell you which one you are holding. This guide covers what each method measures, what real-world data says about normal capacity loss, where reports mislead, and why the warranty is a separate question. Facts below are as of November 2026.

Key takeaways

Ask three things of any battery report: when it was produced, how (test, data estimate or projection), and whether it covers the exact VIN you are buying.
Geotab’s telematics study of about 22,700 EVs found average capacity loss near 2.3% a year, but charging habits and climate shift that, and individual cars vary.
A screenshot is one reading on one day, usually supplied by the person selling. A record over months is harder to dismiss.
Recurrent’s dealer tools project range from fleet data and say they do not measure battery health directly.
The warranty and the report answer different questions. Get the terms for your VIN in writing, including the capacity threshold and whether coverage passes to a second owner.

Three kinds of “battery health” number

State of health (SoH) generally means the share of a battery’s original usable capacity that is left. How you arrive at that share is where reports differ.

Method	Example from the sources	How it gets its number	Main limit
Direct test	Aviloo’s Flash Test (per electrive, September 2, 2026)	Reads the battery through the car’s OBD interface to produce a manufacturer-independent SoH	Aviloo sells the test; electrive gives no detail on how the 500,000-plus tested cars were selected
Telematics estimate	Geotab	Compares energy taken in while charging with energy used while driving, then tracks how state of charge changes over time	Fleet averages; trends, not guarantees; it does not examine chemistry or battery-management design
Statistical range projection	Recurrent’s dealer tools	Expected range now and later, modeled from anonymized data on more than 50,000 EVs	Recurrent says it does not measure battery health directly and that without a lab test, health can’t be guaranteed

Aviloo also claims that the battery displays built into cars are often unreliable. That is a vendor’s claim, and the electrive article does not show the evidence, so treat it as a reason to ask, not a finding. What is safe to say: the range on your dashboard is an estimate of distance, and Recurrent notes that driving habits, climate, location and charging patterns all move real-world range.

What normal capacity loss looks like

Geotab’s updated guide, dated April 28, 2026, uses telematics from about 22,700 EVs across 21 make-models. Its headline: roughly 2.3% capacity loss per year, projecting to about 81.6% of original capacity after eight years.

Group (Geotab)	Approximate capacity loss per year
All vehicles in the study	2.3%
DC fast charging in 12% of sessions or fewer	1.5%
DC fast charging in more than 12% of sessions	2.5%
High fast-charging users with over 40% of fast sessions above 100 kW	3.0% (versus 2.2% for those below that level)
Hot climates (over 35% of days above 25°C, or 77°F)	About 0.4 points a year faster than mild climates

Geotab stresses what these numbers are not. They are group averages, so an individual car can lose more or less. The study did not analyze battery chemistry or design, had too few cold-climate-only vehicles to measure extreme cold, and charging behavior correlates with other factors. Read it as a baseline for expectations.

The spread inside a single model is the more useful lesson. In the Aviloo data described by electrive, the Hyundai Ioniq 5 (72.6 kWh) averaged 92.8% SoH at 150,000 km (about 93,000 miles), yet individual cars differed by more than 12 percentage points, roughly 42 km (26 miles) of range. Aviloo’s medians also fell as mileage rose and the gaps between cars widened. Same model, same mileage, different batteries. That is the argument for testing the specific car.

Where a battery report can mislead you

It is one reading, usually from the seller

Recurrent’s September 19, 2026 newsletter describes a Buffalo-area dealer who asks sellers for a battery test and a screenshot before making an appraisal offer. He says nearly all cars pass, around 99% of the time, because the team grades against thresholds for model year and mileage rather than perfection. Recurrent itself notes the limits: a screenshot from the morning of an appraisal is a single reading, the seller who wants a higher offer is the one producing it, and several months of history compared with similar cars is harder to wave away.

The vendor has a business

Recurrent sells battery-tracking services, Aviloo sells tests and Geotab sells telematics. That does not make their numbers wrong, but it is context. Recurrent says its data covers about 1 billion miles of EV driving and that only about 4% of EV batteries have failed in its research. Both are the company’s own figures, not an independent audit.

Percentages are not portable

A percentage from a direct test, a figure from a range projection and a number on the dashboard are three different statements. Recurrent says its “expected range” reflects what it observes when a vehicle is new, which can differ from EPA figures. Do not compare scores across services unless the method matches.

The warranty is a separate question

A report describes the battery’s condition. A warranty sets the terms under which a manufacturer may fix or replace it. Ford’s page “What is the Ford warranty on batteries?” says EV and hybrid batteries are covered for 8 years or 100,000 miles from the warranty start date, whichever comes first, and that coverage includes excessive capacity loss. It does not state the capacity threshold, does not address whether coverage transfers, and points to other pages for the start date and exclusions. If a manufacturer’s own FAQ leaves those out, a seller’s summary will too.

Arithmetic matters here. Take a hypothetical 8-year, 100,000-mile term, whichever comes first. A car bought at 4 years old with 60,000 miles has 4 years and 40,000 miles left on paper. Drive 15,000 miles a year and the mileage runs out in about 2.7 years. I could not open several other manufacturers’ warranty pages while researching this, so I am not quoting their terms. They vary by brand and model year, so read the warranty booklet for the VIN.

Frequently asked questions

Is a screenshot from the seller good enough?

It is a starting point. Ask when it was taken, what tool or method produced it, and whether it matches the VIN. A report you commission or a record over several months is stronger evidence.

How much battery loss is normal?

Geotab’s average is about 2.3% per year, with fast-charging habits and hot climates pushing it up. Individual cars vary, as the Ioniq 5 spread above shows.

Does the dashboard range tell me the battery’s health?

Not directly. It estimates distance and moves with driving, climate and charging patterns. Use it as a clue, then ask for a method-backed report.

Will the manufacturer warranty cover capacity loss?

It may, but terms differ. Ford’s page lists capacity loss within an 8-year, 100,000-mile term without giving a threshold. Check the threshold, start date and transferability for your car.

Before you decide

This guide does not recommend any vehicle. It is a list of questions to settle before you rely on a number. A reassuring report also does not create a convenient charger at home or shorten a long commute, so compare the car with your real routes and charging access alongside the inspection and service history.

Date, method and VIN of the report: test, data estimate or projection?
One reading or months of history? Who produced it and who paid?
Can you get an independent test or inspection of your choosing?
Warranty in writing: years, miles, start date, capacity threshold, second-owner transfer.
Service records and charging history: how often fast charging, and which climate.
Fit with your driving: daily distance, charger access at home or work.