Buying a used electric car: how to judge the battery's health

Range anxiety drives used-EV pricing more than any other single factor, yet most of that fear rests on anecdote: a friend's phone-battery experience scaled up to a car, or a horror story that travelled further than it deserved. We aggregated state-of-health (SoH) telemetry from thousands of used electric vehicles to replace guesswork with a distribution you can actually reason about.

Degradation is front-loaded, not linear

The figure most buyers fixate on, capacity lost per year, is misleading because the loss is concentrated early. Our dataset shows an average 3.1% drop in the first year that settles to roughly 1.8% annually thereafter, as cell chemistry stabilises and the pack settles into its working life.

This shape matters enormously for buyers. A three-year-old EV has already taken its steepest hit, so a used buyer inherits the flat part of the curve rather than the cliff, much as a combustion buyer benefits from someone else absorbing the first wave of depreciation.

• Year 1: 3.1% average capacity loss, the break-in phase.
• Years 2-5: 1.7-1.9% per year, the stable plateau.
• Years 6-8: a mild reacceleration to 2.2% as cycle counts climb.
• Median fleet SoH at 8 years: 87.4%.

What actually accelerates degradation

When we regress SoH against usage variables, three signals dominate the model. Together they explain roughly 71% of the variance in degradation rate, which means the rest is closer to luck, climate and cell-batch lottery than most listings would have you believe.

1. DC fast-charge share: vehicles charged on DC more than 40% of the time degrade about 1.6x faster.
2. Thermal exposure: sustained operation in hot climates adds measurable annual loss.
3. Depth of discharge: routinely cycling from full to near-empty stresses cells more than shallow top-ups.

What is striking is how much these are behaviours rather than fixed traits. Two identical cars delivered the same week can diverge by several points of SoH within a few years purely on how, where and how aggressively they were charged.

The number that matters at resale

A SoH reading on its own is noise without context. We pair it with the model's expected curve so a 91% reading on a four-year-old car reads as ahead of schedule rather than as a bare percentage. That comparison is what protects you from overpaying for a tired pack dressed up in a clean body.

An EV's body can be detailed to perfection; its battery cannot be polished. The pack is the part of the car that tells the truth.

Why we resist a single headline figure

A confident wrong number costs more than an honest range. Where telemetry is sparse we widen the confidence band rather than inventing precision, because an EV bought on a falsely exact SoH can disappoint by the kilometre for years afterward.

Why the warranty is not the whole story

Buyers often lean on the manufacturer battery warranty as a safety net, but its threshold is usually low, commonly a guarantee against dropping below 70% capacity within a set term. A pack can degrade well beyond what an owner would accept and still sit comfortably inside warranty, leaving the buyer technically covered but practically short on range. The SoH reading, not the warranty line, is what tells you how the car will actually feel to live with.

Reading the signal on a report

On every EV report we surface the SoH estimate, its confidence band, and the charging mix behind it, so you can see not just where the pack stands but why it got there. A car with a high SoH and a gentle charging history is a genuinely different proposition from the same SoH reached through heavy fast-charging that may steepen the decline ahead.

Used responsibly, this turns the most feared part of an EV purchase into the most measurable. The battery stops being a black box and becomes the clearest data point on the report.