Most Plane Crash Deaths in America Don’t Happen on Airlines

The footage always arrives before the facts do. A column of smoke over a treeline, a debris field taped off in a field or a river, and a casualty number that climbs through the evening as the networks cut back to it.

For three or four days the crash is the only aviation story in the world. Then it fades, and at some point you remember you have a flight booked next month.

So how often does this actually happen? There is a real number, and it is smaller than the coverage implies. But the number that matters more is a second one almost nobody quotes, because your risk depends far less on flying than on which kind of aircraft you are getting into.

How Often Do Planes Crash?

In 2025 the world’s commercial airlines operated 38.7 million flights and recorded 51 accidents of any severity. That works out to 1.32 accidents per million flights, or one for every 759,646 departures, according to the IATA Annual Safety Report published in March 2026.

Most of those 51 were survivable. Eight were fatal, killing 394 people on board. Measured across the five years from 2021 to 2025, that is one fatal accident for every 5.6 million flights.

It helps to know that “accident” here is a defined technical term, not a synonym for a fireball. It covers any event involving a fatality, a serious injury, or substantial damage to the aircraft, which includes runway excursions and hard landings that everyone walks away from.

Where the Rate Has Actually Gone

Year-to-year movement in these figures is mostly noise. The trend underneath it is not.

Measure	2024	2025	2021-2025 average
Accidents (all severities)	54	51	44
Fatal accidents	7	8	6
Onboard fatalities	244	394	198
Accident rate per million flights	1.42	1.32	1.27
Fatality risk per million flights	0.06	0.17	0.12

In 2005 the global accident rate was 3.72 per million flights. Twenty years later it is 1.32, roughly a third of what it was.

Fatal accidents specifically have thinned out even faster. Between 2012 and 2016 there was one for every 3.5 million flights. Between 2021 and 2025 it took 5.6 million flights to produce one.

None of that happened on its own. Almost every step down in that curve traces back to a specific investigation and the rule, checklist, or piece of hardware that came out of it, which is the whole story of the crashes that rewrote how aviation works.

Why 2025 Felt Worse Than the Trend Line

Here is the tension in the 2025 data. There were fewer accidents than in 2024, yet 61 percent more people died on board.

That is what a small number of very bad outcomes does to an otherwise improving year. Two accidents accounted for more than 77 percent of all onboard deaths in 2025.

The first was Air India 171, a Boeing 787 that came down shortly after takeoff from Ahmedabad in June 2025, killing 241 of the 242 people on board and 19 more on the ground. The second was PSA Airlines 5342, a CRJ700 that collided with a US Army Black Hawk over the Potomac River on the night of 29 January 2025, killing all 64 on the aircraft and all three on the helicopter.

The Potomac collision landed harder in the United States than the raw number explains, because of the streak it broke. The previous fatal crash of a US scheduled airline was Colgan Air 3407 outside Buffalo in February 2009.

A run that clean is not the historical norm, and it is not a baseline anyone can hold forever. When a nearly 16-year gap ends, ordinary statistical variation reads to the public like a system falling apart.

It is also why the accidents people remember by name are so heavily weighted toward a handful of years. Rare events do not distribute themselves politely across the calendar.

Almost All of the Risk Sits Outside the Airlines

This is the part that reframes the question. In 2024 there were 1,201 accidents involving US-registered civil aircraft and 321 deaths, on NTSB preliminary figures.

Not one of those 321 people died on a scheduled US airline. Virtually all of them died in general aviation: privately owned piston singles, flight training, personal cross-country trips, agricultural work, air taxi and backcountry flying.

That pattern is not a one-year fluke. Across the last decade, US civil aviation has produced more than 1,100 accidents and over 300 deaths in a typical year, and the overwhelming majority of both sit in general aviation rather than at the airlines.

The rate tells the same story. The FAA put the general aviation fatal accident rate at 0.68 per 100,000 flight hours in fiscal 2024, its lowest since the agency began tracking it in 2009, off 195 fatal accidents and 337 fatalities. The comparable figure for scheduled airlines that year was zero.

The reason is structural, not a matter of luck. A light aircraft is usually flown by one pilot with one engine, no dispatcher, no relief crew, no second opinion on a marginal weather call, and often into a short strip with no approach lighting and no control tower.

An airline flight has two qualified pilots, a dispatcher sharing legal responsibility for the plan, redundant systems on nearly everything that matters, mandatory rest rules, and a maintenance regime that is audited. Those layers are the entire reason airline flying is as safe as it is, and general aviation has almost none of them.

So a headline crash statistic is close to meaningless until somebody names the operation. “How often do planes crash” and “how often do airliners crash” are questions with answers two orders of magnitude apart.

The Myth: Crashes Are Getting More Common

Three things drive the illusion, and they compound.

The first is coverage. Every incident now arrives with phone video from the ground, a public flight-tracking trace anyone can replay, and days of expert speculation before the investigators have finished walking the site.

The second is that absolute counts and rates move differently. Global flight volume rose from 37.9 million in 2024 to 38.7 million in 2025, so a flat number of accidents against a growing denominator is actually an improving safety record.

The third is clustering. Genuinely random rare events arrive in clumps rather than evenly spaced, so two serious accidents in the same month feel like a pattern when they are the statistical equivalent of two heads in a row.

The same distortion applies to the hazard passengers actually fear most in the cabin. Severe turbulence injures a meaningful number of people every year, mostly unbelted crew and passengers, but it is almost never what brings an aircraft down.

There is one place the averages genuinely mislead, though, and it is geography. In 2025 the accident rate in Africa was 7.86 per million flights, while North Asia recorded 0.16, a spread of nearly fifty to one.

Whether that gap touches you depends on operator, aircraft type, and region rather than on flying as an activity. The global average is a useful headline and a poor personal risk estimate.

A Note on the Driving Comparison

The usual reassurance is that driving is more dangerous, and on absolute numbers it is not close. The US recorded an estimated 39,345 road deaths in 2024 against 321 civil aviation deaths, and again, none of the aviation deaths were on a scheduled airline.

The comparison is worth handling honestly, because the measure you pick changes the size of the gap. Per mile travelled, aviation looks extraordinarily good, since a single flight covers a distance most drivers take a week to accumulate.

Per journey the gap narrows, because takeoff and landing carry most of the risk in a flight regardless of how far it goes. Airlines still come out ahead on that measure, just by less than the per-mile figure suggests.

That distinction matters if you fly short hops constantly. It does not change the conclusion for the once-or-twice-a-year passenger, whose annual exposure is a rounding error either way.

Next time the coverage runs for three days, two numbers are worth keeping in your head. One fatal airline accident for every 5.6 million flights, and zero scheduled US airline deaths in 2024 out of 321 aviation deaths that year.

The second number is the one that answers the question people are really asking. The risk in aviation is real, it just is not distributed anything like the way the news makes it look, and almost none of it is sitting in seat 24B.