At What Mileage Do Cars Actually Fail? We Checked 1,110,092 NHTSA Complaints — and the Obvious Answer Was Wrong

Key facts

• 62,540 miles — median odometer reading at failure across all retained complaints (n = 1,110,092).
• 29.4% of complaints report the failure under 36,000 miles, inside a typical bumper-to-bumper warranty.
• 25.3% report it at 100,000 miles or more.
• 51.3% of vehicle complaints supply a mileage figure at all; the other 1,022,858 were excluded.
• Source: NHTSA ODI complaint flat file, last-modified 2026-08-20 01:32 GMT, retrieved the same day.

At what mileage do cars start having problems?

There is no single answer, and that is the finding. The median complaint in NHTSA's file reports a failure at 62,540 miles, but the interquartile range runs from 29,600 to 100,000 miles. Nearly three in ten reported failures happen before 36,000 miles — while most cars are still under original bumper-to-bumper coverage — and a quarter happen past 100,000. Any advice that names a single "problems start here" mileage is discarding most of the distribution.

This matters right now because affordability is pushing American buyers up the odometer. With the average three-year-old used car above $32,000, shoppers who once bought lightly-used are moving into older, higher-mileage inventory. The useful question is not "how many miles is too many" but "what do this car's own records say."

Which components fail earliest, and which fail latest?

Ranking the 35 component groups with at least 500 complaints, the earliest and latest medians look like this. Read the second column with the caution in the next section — it is the one that broke.

The check that reversed our headline

The obvious story in that table is that driver-assistance electronics fail almost immediately — forward collision avoidance at 16,890 miles against 74,000 for engines, a 4x gap. It is also wrong, and the reason is simple: forward collision avoidance only exists on newer cars, and newer cars have not driven many miles yet.

The confound is visible directly in the data. Median complaint mileage falls monotonically with model year, from 75,000 miles for model-year 2000 vehicles to 5,000 miles for model-year 2025 — not because new cars fail sooner, but because a 2025 car cannot yet report a failure at 80,000 miles.

So we recomputed the entire ranking restricted to model years 2015-2018, where every vehicle has had at least eight model years to accumulate mileage and ADAS is already common. The ranking reshuffled:

Forward collision avoidance more than doubled and moved from dead last to mid-pack. Engines moved the other way. The "new tech fails first" reading does not survive the control, and we are reporting that rather than publishing the more quotable version. What survives is the wear ordering inside the cohort: tires (17,560) and wheels (18,000) earliest, suspension (25,000) and brakes (29,000) next, engine cooling (86,000) last.

What the fresh Stellantis recall does and doesn't illustrate

On 17 August 2026 Stellantis recalled 848,511 US vehicles — 2026 and 2027 Chrysler Pacifica, Voyager, Dodge Charger, Jeep Cherokee, Compass, Gladiator, Grand Cherokee, Grand Wagoneer, Wrangler and Ram 1500, 2500 and ProMaster models — because radio software can stop the rear-view camera image appearing on the media screen. The remedy is an over-the-air update.

It is tempting to file this under "modern electronics break early." Our own data says be careful with that inference: those are near-new vehicles, so a low failure mileage is what you would expect regardless of whether the electronics are unusually fragile. A recall on a 2027 model year tells you about that campaign, not about how software ages.

By the numbers

Method: how we counted, so you can reproduce it

We used the NHTSA Office of Defects Investigation consumer complaint flat file, FLAT_CMPL.zip, source file last-modified 2026-08-20 01:32 GMT and retrieved on 2026-08-20. The file is public and requires no key.

• Universe. All 2,236,359 records; 2,164,355 remained after keeping only PRODTYPE = 'V' (vehicle) rows.
• Filter. We required the MILES field — position 18 in NHTSA's CMPL record layout, the odometer reading the complainant reported at the time of failure — to parse as an integer between 1 and 500,000. Zero and blank mean "not supplied", not "failed at zero miles", so both were excluded rather than counted as low mileage. That left 1,110,092 records, 51.3% of vehicle rows.
• Cleanup. Records whose make contained TEST or TBD, or whose model year was 0000/9999/blank, were dropped (15,032).
• Dedup. Identical resubmissions were collapsed on ODINO + make + model + model year + full component description + mileage (8 records).
• Aggregation. Records were grouped by the top-level segment of COMPDESC (the text before the first colon), and we report the 25th, 50th and 75th percentiles of MILES per group. Groups with fewer than 500 records are suppressed, leaving 35 groups.
• Fleet-age control. The identical aggregation was rerun restricted to model years 2015-2018, and both rankings are published side by side in the CSV.

Known limits. Complaints are voluntary and unverified. Mileage is self-reported and heavily rounded (27.6% are multiples of 5,000). Roughly half of complaints supply no mileage at all, and we cannot rule out that those differ systematically from the ones that do. Most importantly, this file has no denominator — no count of vehicles on the road — so it supports statements about when owners report failures, never about failure rates, reliability rankings, or which brand is worst. Complaint volume also spikes with recall publicity.

Frozen dataset: nhtsa-mileage-at-failure-by-component-2026-08-20.csv

What a VIN check can and can't tell you here

If the odometer number is the axis everything turns on, the number itself is worth verifying.

A Zilocar VIN check can show:

• Recorded odometer readings from past sales, listings and inspection events, and rollback flags where a later reading is lower than an earlier one
• Accident and damage records, including airbag-deployment events
• Junk and salvage auction records
• Theft records (NICB)
• Ownership history and past sales-listing history
• Whether the VIN is named in an open safety recall, plus specs, NHTSA and IIHS ratings, and a valuation

It cannot show:

• Whether a recall has actually been remedied or the software flashed — for the Stellantis campaign above, only a dealer or NHTSA's own VIN lookup confirms remedy status
• Open NHTSA investigations (PE/EA), which are not recalls
• Per-unit firmware or module versions
• A legal title-brand classification

For remedy status on any specific VIN, use the free NHTSA recall lookup or ask a franchise dealer.

Related checks

• Odometer rollback: how to spot it by VIN
• What recall status versus investigation means on a VIN check
• Which used cars have the most backup-camera failure complaints

Sources

• NHTSA ODI consumer complaint flat file, FLAT_CMPL.zip — https://static.nhtsa.gov/odi/ffdd/cmpl/FLAT\_CMPL.zip — source file last-modified 2026-08-20 01:32 GMT. Last verified 2026-08-20.
• NHTSA recall lookup and recall database — https://www.nhtsa.gov/recalls — Last verified 2026-08-20.
• Stellantis rear-view camera recall, 848,511 US vehicles, announced 17 August 2026, as reported by CBS News and CBT News. Last verified 2026-08-20.
• Zilocar sample vehicle history report — https://zilocar.com/vehicle-history-report-example — Last verified 2026-08-20.