U.S. Airlines Had 37 Accidents in the First Ten Months of 2025, the Most Since at Least 2008. The Count for 2026 Is 20.

The National Transportation Safety Board keeps a public database of every civil aviation accident it investigates in the United States. The version posted on 1 November 2026 lists 20 accidents involving U.S. airline flights between 1 January and 28 October this year. For the same ten months of 2025 it lists 37, the highest figure for that stretch in any year since the file begins in 2008.

So is American air travel getting less safe? On the board's own records, 2025 was the worst year for airline accidents since at least 2008, but 2026 is running below the long-run norm: 20 accidents to the end of October against an average of 23.4 for 2008 to 2024, and no deaths aboard an airliner.

That is a count, and counts of rare events move around. The more useful questions are what the accidents were, and whether last year's jump is the start of something or a bad year that has passed.

What counts as an airline accident?

Less than most readers would guess. Federal rules define an aircraft accident as an occurrence between boarding and disembarking "in which any person suffers death or serious injury, or in which the aircraft receives substantial damage," according to 49 CFR 830.2. A serious injury includes a hospital stay of more than 48 hours or "a fracture of any bone (except simple fractures of fingers, toes, or nose)."

That threshold means a flight attendant who breaks an ankle in turbulence produces an accident report, as does a wingtip clipping another aircraft on a taxiway. Both appear in the same column as a crash. We counted every event the NTSB classes as an accident where at least one aircraft was operating under Part 121, the rule book for scheduled airlines and large cargo carriers, and sorted each by the defining event the board recorded.

How do the years compare?

2025 stands out; 2026 does not. The table sets each year's January-to-October count beside the turbulence share, the deaths and a rate that allows for the growth in flying.

The passenger rate uses enplanement data from the Bureau of Transportation Statistics, which count every passenger boarding on U.S. carriers' scheduled flights. U.S. airlines boarded 565.5 million passengers from January to July 2026, against 539.4 million in the same months of 2019. More flying means more chances for something to go wrong, so a flat accident count would itself be an improvement. By that measure, the first seven months of 2026 were the second-safest in the file after 2022, a year when the count was unusually low for reasons the data do not explain.

What made 2025 so bad?

Two things: one catastrophe and a run of in-flight injuries. On 29 January 2025, PSA Airlines flight 5342, a CRJ700 regional jet, collided with a U.S. Army Black Hawk helicopter about half a mile southeast of Reagan Washington National Airport. All 64 people aboard the jet and the three soldiers aboard the helicopter died, according to the NTSB's investigation page. Those 67 deaths are every airline-accident death in the first ten months of 2025. The board has since found the probable cause to include "the Federal Aviation Administration's (FAA) placement of a helicopter route in close proximity to a runway approach path."

The rest of the year was turbulence and cabin injuries. Sixteen of the 37 accidents were turbulence encounters, and the database records 25 serious injuries in those ten months. Southwest Airlines alone appears in eight accidents in January to October 2025, almost all involving one seriously injured person aboard.

Then, just outside the ten-month window, UPS flight 2976, an MD-11F freighter, crashed shortly after takeoff from Louisville on 4 November 2025. The three crew members and 11 people on the ground died, the NTSB said.

Was 2025 a trend or a bad year?

On this evidence, a bad year. A rough test: if airline accidents arrive at random around the 2008-2024 average of 23.4 for January to October, the typical year-to-year spread is about five accidents either way. Thirty-seven sits almost three of those spreads above the average, which makes 2025 an unusual year rather than ordinary noise. Twenty sits comfortably inside the normal range. That is our arithmetic, using a standard model for rare events, not a finding by the NTSB.

Two details in 2026 deserve attention all the same. Cargo flights account for 5 of the 20 accidents, the most for the period since 2008, when there were six. And the one fatal accident was a cargo flight. A 21 Air Boeing 767 in Amazon Prime Air livery overran runway 30 at Miami on 6 September and killed five people on the ground. The NTSB's investigative update of 9 September said the nose and right main gear touched down at a groundspeed of 158 knots, and that a pilot called for a go-around about 15 seconds before the cockpit recording ended. We reported then on who held the operating certificate and on the warnings about speed captured on the recorder.

What the database does not show

It does not count close calls. An incident, such as a runway incursion where no one is hurt and nothing is damaged, sits in a separate category and is investigated far less often. Many of the questions raised since the Washington collision concern exactly that kind of event, and the strain on air traffic control staffing that sits behind them. A year can have few accidents and many near misses.

The database is also provisional at the recent end. Investigations take a year or more, and the codes on 2026 entries are preliminary. Accidents abroad investigated by foreign authorities, such as the Endeavor Air jet that overturned on landing at Toronto in February 2025, carry fewer details. We counted them where the NTSB lists them.

What to watch

The NTSB's monthly update files to the database are dated the first of each month, so the next is due on 1 December, the same day the St. Louis Fed lists for the next month of BTS passenger numbers. The busiest flying of the year falls between those updates and the end of December. If the 2026 count finishes near 25, it will be an ordinary year by the standard of the past two decades. The final report on the Miami overrun, likely a year or more away, will say whether this year's one fatal crash was about a single operator or something wider.

Accident counts, dates, operators, injury levels, fatalities and defining events are from the National Transportation Safety Board's aviation accident database (avall.mdb), downloaded on 13 November 2026; the file is dated 1 November 2026 and its latest event is dated 28 October 2026. Passenger enplanements for U.S. carriers' scheduled domestic and international flights are from the Bureau of Transportation Statistics, as published by the Federal Reserve Bank of St. Louis (series ENPLANE), with data to July 2026. The definitions of an accident and a serious injury are from 49 CFR 830.2. The Miami overrun details are from the NTSB's investigative update of 9 September 2026. The counts, the classification by type, the rates per 10 million enplanements and the variance arithmetic are this publication's. Recent entries carry preliminary codes and may be revised. The analysis is our own.

Topics businessaviationairlinesNTSBeconomic data