Every NHTSA recall has a remedy field describing the fix. Reading it across all 12,666 campaigns since 2018 shows the industry quietly changed what a recall means.
| Brand | Fixed by software | |
|---|---|---|
| Tesla | 74% | 418 campaigns |
| Volvo | 44% | |
| Mazda | 44% | |
| Jaguar | 41% | |
| Market average | 27.2% | |
| Mercedes-Benz | 32% | |
| Kia | 31% |
A recall used to mean a day without your car
In the 1990s, 0.6% of recalls were fixed with software. In the 2020s it's 29%. Over the same period the share fixed by replacing a physical part actually fell slightly, from 62.6% to 56.1%, while "inspect" jumped from 22.8% to 33.4%.
The phrase "over-the-air" appears in 1,036 remedy descriptions, rising from 15 mentions in 2014 to 149 in 2021.
Why Tesla is an outlier and what it costs
74% is not three times better engineering — it's a different architecture. When most vehicle functions are controlled by software, most defects are software defects, and they're patched remotely.
The upside for an owner is real: no appointment, no loaner, no day off work. The car updates in the driveway.
The downside is that a recall stops feeling like an event. A campaign that arrives silently overnight doesn't register as "my car had a safety defect" — and it still shows up in the vehicle's history exactly like one that required a new part.
What this means when you're buying used
Recall counts aren't comparable across brands any more. Ten Tesla recalls and ten Kia recalls describe different amounts of disruption and, often, different severity. Counting campaigns without reading them is now actively misleading.
Software recalls can be outstanding without anyone noticing. If the car wasn't connected, or the owner declined the update, the campaign stays open. There's no rattle to prompt a visit — the same problem we found across the board when we checked what's open on cars our readers own.
Ask whether updates were actually installed. On a connected car this is checkable in the vehicle's own menus, and a dealer can confirm the software level against the campaign.
Read the remedy, not the number. A VIN check lists the open campaigns for a specific vehicle so you can see what each one actually requires — a download, an inspection or a part. We looked at how the subject of recalls changed in from headlights to cameras, and at how long each type takes to surface in which recalls take longest.
Frequently asked questions
How many recalls are fixed with a software update?
27.2% of campaigns since 2018 across the market, up from 0.6% in the 1990s. Tesla leads at 74%, Volvo and Mazda at 44%, while Kia sits at 31% and Mercedes-Benz at 32%.
Why does Tesla have so many software recalls?
Architecture, not engineering quality. When most vehicle functions are controlled by software, most defects are software defects — and they're patched remotely rather than at a workshop.
Is a software recall less serious?
Not necessarily. It's less disruptive to fix, but it appears in the vehicle's history the same as any other campaign and addresses a defect NHTSA considered a safety risk.
Can a software recall stay unrepaired?
Yes. If the car wasn't connected or the owner declined the update, the campaign remains open — and unlike a mechanical fault there's no symptom to prompt a service visit.
Denis Kataev is a serial solopreneur and the founder of ForCar. With 15+ years in software engineering and 10 years in SEO, he builds data-driven products end to end — backed by a sharp eye for design. At ForCar he mines proprietary vehicle datasets, turning raw numbers into buying advice you can actually trust.
Keep reading
All articles →
An Airbag Inflator Defect Takes 9 Years to Be Admitted. A Seat Belt Fault Takes 2.
Some defects get found almost immediately and some hide for a decade. Sorting recalls by how long they took to surface produces a ranking that says a lot about which parts fail slowly.
A Clean Recall Check Today Means Nothing. A Third of Recalls Arrive Five Years Late.
Recall numbers carry a hidden date stamp. I used it to measure how long a defect stays undiscovered — and the answer changes what a clean report is actually worth.
The Heaviest Vehicle on the Road Weighs as Much as Five of the Lightest
Two vehicles sharing the same lane can differ in mass by a factor of five. Star ratings never account for it, because cars are only ever tested against others their own size.