“Electric cars catch fire more often than combustion cars”

What is actually being claimed?

A burning car proves that a fire occurred. It does not establish how often fires occur in that powertrain. The comparative claim needs both a count and an exposure denominator: vehicles in service, vehicle-years or distance travelled. A short compilation of dramatic incidents supplies neither. Our verdict concerns the blanket frequency claim; it does not imply that battery fires are harmless or that every country has the same risk.

The comparison also needs to separate an incident that starts in a vehicle from one that spreads to a parked vehicle from elsewhere. Without a consistent incident definition, two published totals may count different events. This is a reason to ask for methodology, not to dismiss inconvenient reports.

The recent evidence

The Danish Emergency Management Agency’s July 2026 factsheet updates the observation year to 2025. It counted 62 dispatches to fires involving electric and hybrid cars, up from 50 in 2024, while the combined registered fleet grew from about 485,000 to nearly 693,000. The agency reports 1.0 fire per 10,000 electric and hybrid cars in 2025, versus 4.0 for other car types; the corresponding combined-category rate was 1.2 in 2024. These are official fleet-based fire frequencies, not a count of battery-initiated fires or a measure per kilometre travelled. ^[4]

NIST’s March 2026 US report combined fire-service records with TeslaFires.com reports and used capture–recapture methods to estimate vehicle-fire counts for 2011–2023. It cautions that EV and plug-in-hybrid fires are poorly identified in NFIRS and that unusually high overlap in Florida records likely biases the national estimates. The resulting counts are model-based, not exposure-adjusted fire rates or a direct comparison with combustion vehicles. ^[5]

Why the comparison has limits

This dataset groups electric cars with hybrids. It cannot isolate the risk of a battery-only car. An overall vehicle fire is also different from a fire that began in the traction battery. Neither chart bar adjusts for fleet age or mileage. Those are reasons to keep the result geographically and statistically scoped, rather than convert the ratio into a universal claim that every EV is a fixed number of times safer.

The Danish agency also warns that Nordic fire totals are not directly comparable: Sweden and Norway may not count a fire that spreads from an electric car to a building as an electric-car fire. For a stronger causal comparison, researchers would ideally use similarly aged vehicles over the same period, consistent incident reporting and equivalent exposure. ^[4]

A useful chart title therefore says reported vehicle fires, Denmark, 2025, and electric plus hybrid combined. It should not quietly change the label to battery fires or all-electric cars. The published rate is fleet-based, not adjusted for vehicle age or distance travelled. The 2025 factsheet gives the registered fleet at year end but does not spell out the precise within-year exposure denominator used for its rate, so the official reported rate should be quoted as published rather than reverse-engineered from the year-end stock count.

A lower frequency does not erase a difficult hazard

NHTSA’s 2024 safety-rule proposal discusses thermal runaway and propagation between lithium-ion cells. Its current safety guidance warns that damaged high-voltage batteries can release hazardous gases and cause immediate or delayed fire. These are credible hazard descriptions, not estimates of how often the average owner will experience a fire. Frequency and consequences answer different questions and should appear separately in a safety discussion. ^[2][3]

How to read the next viral fire story

Check whether the vehicle was battery electric, plug-in hybrid or conventional hybrid; whether the battery initiated the incident; and whether the source reports all fires or selected cases. Then ask what denominator and comparison period were used. If those details are absent, classify the comparative claim as unestablished by that story. Use official recalls and vehicle-specific safety guidance for a particular model, because a category average cannot diagnose an individual defect.