Speeding Killed 11% Fewer People in 2025. Speed Limits Didn't Change. The Cars Did.

NHTSA's latest subcategory report, published this month, contains a number that deserved front pages but didn't get them: speeding-related traffic fatalities fell 11% from 2024 to 2025, outpacing the overall 6.7% decline in road deaths and beating improvements in every other subcategory except ejections and unrestrained occupants.^[1] For a country that treats triple-digit highway speeds as a constitutional right, that sounds like progress.

Nobody lowered a speed limit to earn it. No state reduced its posted maximums, and several actually raised them on rural stretches where the killing is worst. No federal mandate for Intelligent Speed Assistance arrived, even though Europe required it starting July 2024. None of that 11% came from regulation, because it came from crumple zones, AEB calibrations, and side-curtain airbags absorbing crashes that would have produced body bags a decade ago.

Buried in the same NHTSA report is the mechanism that explains how: occupant fatalities in newer passenger vehicles, those under 10 years old, fell 12%, while fatalities in older vehicles dropped just 9%.^[1] A three-percentage-point gap between newer and older fleets is the fingerprint of engineering, not enforcement. IIHS quantified the aggregate result in June when it estimated that crashworthiness improvements alone have saved approximately 50,000 lives since 1995, worth roughly $500 billion in societal cost savings.^[2]

FARS data makes this gap visceral at the model level: a Ford Mustang accumulates 6.02 deaths per 100 million VMT while a Toyota RAV4 accumulates 0.19, producing a 31.7-to-1 ratio between two vehicles sharing the same roads and the same speed limits.^[3] Even within GM's own sports car lineup, a Corvette (1.52) kills at less than half the rate of a Camaro (3.44), despite both carrying nearly identical impairment profiles and marketing to the same buyer demographic. Same brand, separated by $30,000 in aluminum architecture and hydroformed frame rails.

Maybe Americans just drove more carefully in 2025 because of post-pandemic normalization, gas prices, or remote work reducing rush-hour aggression? All plausible explanations that fall apart under the data: total vehicle miles traveled increased in 2025, rural interstates where speeds run highest saw fatalities go up 1%, making them the only road category that got worse, and urban arterials dominated by newer commuter vehicles saw the sharpest drop at 12%.^[1] If Americans discovered restraint, they forgot about it on I-80 through Nevada.

What actually happened is the most expensive speed limit in American history: instead of investing $5 billion in speed cameras, road redesign, and ISA mandates, the country spent roughly $50,000 per new vehicle to armor-plate its citizens against velocity they refuse to regulate. Europe reduced per-capita road fatalities to roughly half the U.S. rate by building narrower lanes and enforcing lower limits, while America built wider trucks with better crumple zones and called it progress. Q1 2026 confirms the trajectory holds, with a fatality rate of 0.99 per 100M VMT, the lowest quarterly rate since 2014.^[4]

Eleven percent is a real number, and the engineering behind it deserves credit. But the 65-and-older age group, which drives older vehicles and absorbs crash forces differently regardless of what the crumple zone accomplishes, saw zero percent improvement in 2025.^[1] For them, the technology substitution hasn't arrived, and a speed limit still would have.

If you're shopping: check your target vehicle's IIHS rating at iihs.org/ratings and prioritize current-generation models with standard AEB and electronic stability control. If you're driving something from the mid-2000s at 2025 speeds, your vehicle is the one variable the 11% improvement didn't reach.

Strongest counterargument: Behavioral change may contribute alongside technology, and a three-percentage-point newer-vs-older vehicle gap, while suggestive of a technology effect, does not prove technology explains all 11 percentage points of the decline without controlling for VMT by vehicle age cohort. Post-pandemic normalization, increased enforcement in some metros, and shifting commuting patterns could each play a role that this analysis cannot cleanly isolate.

Limitations: FARS model-level data covers 2014–2023 and cannot directly confirm 2025 dynamics; NHTSA’s “speeding-related” classification involves officer judgment with known reporting lag; and the newer-vs-older vehicle gap includes fleet composition effects beyond technology, such as different driver demographics and usage patterns, that cannot be fully disentangled from this data alone.