Every Road in America Got Safer in 2025. Except One.

NHTSA dropped its 2025 subcategory fatality estimates this month, and the headline numbers are genuinely good: 36,640 people killed, down 6.7% from 2024's 39,254.^[1] Urban arterials shed 12% of their deaths. Urban interstates lost 10%, speeding fatalities fell 11%. Even motorcyclist deaths, that perennial horror show, dropped 8%. Line after line of improvement, the kind of across-the-board progress that makes traffic safety researchers forget their cynicism for five minutes.

Then you get to rural interstates.

One thousand eight hundred and twenty-three dead, up from 1,806 the year before.^[1] A 1% increase while every other road functional class in the country declined. That might not sound like much, seventeen additional bodies, but context turns that number into an indictment: urban interstates, the road class most similar in design to rural interstates, saw deaths plummet 10%, or roughly 321 fewer fatalities across the year, which means the divergence between what happened on high-speed roads inside cities and high-speed roads outside them is a gap you can measure in hundreds of lives.

The full breakdown looks like a class photo where one kid is failing while everyone else made honor roll. Rural local and collector roads dropped 3%, urban locals dropped 4%, and rural arterials eked out a 1% decline, modest but still pointed in the right direction. Rural interstates stood alone, heading the wrong way on a one-way improvement trend.

Why? Start with speed. Rural interstates carry the highest posted limits in the country, routinely 70 to 80 mph depending on the state, and at those velocities the physics of a crash turn from survivable to catastrophic with far less margin than on a 45-mph urban arterial where AEB has time to bleed off kinetic energy before impact. Speeding-related fatalities overall fell 11% in 2025.^[1] But that national decline is heavily weighted by urban improvements. On a road where the posted limit already tolerates 75, "speeding" means 90-plus, and the enforcement infrastructure is sparse enough that nobody stops you.

Then there is the fleet, and the fleet is old. Occupant fatalities in passenger vehicles younger than 10 years dropped 12% in 2025; in vehicles 10 years or older, only 9%.^[1] ADAS works, AEB works, lane departure warnings work, but they work for people who own vehicles equipped with them, and rural drivers disproportionately do not. The average rural household keeps its vehicles longer, buys used more often, and is less likely to have access to the kind of dealership service infrastructure that processes recall repairs in a reasonable timeframe. The safety technology revolution is happening, but it is happening in ZIP codes with traffic lights and Whole Foods, not on I-80 through the Nevada basin.

EMS response time compounds the problem in ways the data cannot fully capture. On an urban interstate, paramedics can reach a crash scene in under 10 minutes; on a rural interstate segment between two county seats 60 miles apart, you might wait 30, and survivable injuries become fatal ones. A ruptured spleen that would have been triaged at a Level I trauma center in 20 minutes instead bleeds out in the back of a volunteer fire department's aging ambulance running code on a two-lane access road. This disparity is invisible in FARS data because FARS only records who died, not who almost did.

The rural-urban gap in overall fatality improvement tells the same story at higher altitude: rural deaths fell 2% while urban deaths fell 10%.^[1] A five-to-one ratio. That is not a difference in driving behavior. It is a difference in infrastructure investment, fleet age, enforcement capacity, and emergency medicine access, four variables that compound each other in rural counties and all of which are trending in the wrong direction relative to their urban counterparts.

Two other categories also bucked the 2025 improvement trend, for what it is worth. Pedalcyclist fatalities climbed 4%, from 1,103 to 1,148.^[1] Adults 65 and older held perfectly flat at 8,039 dead.^[1] Zero percent change. Every age cohort between 15 and 64 improved; the 25-to-34 bracket dropped 13%. The pattern repeats: if you are young enough to be driving a recent-model-year vehicle with current safety tech on a well-maintained urban road, 2025 was measurably safer than 2024. If you are on a bicycle, over 65, or driving an older vehicle on a rural interstate at 75 mph with the nearest trauma center 40 miles away, the national progress report does not apply to you.

None of this should surprise anyone who has driven I-10 through rural Texas, I-90 through Montana, or I-70 through eastern Utah. Those roads have not changed in decades, and that is precisely the point. The pavement is the same, the shoulders are the same width, the median barriers are the same age, the cell coverage is the same quality, and the nearest hospital is the same distance away. What changed in 2025 is that every other road category in America improved around them, leaving rural interstates as the last holdout, the one line on the chart still pointing up while everything else bends down.

Seventeen extra dead people. Against a backdrop of 2,614 fewer dead people everywhere else. That is not a rounding error. It is a policy choice.

Source: NHTSA FARS 2024 ARF and 2025 statistical projections (DOT HS 813 829, July 2026). Subcategory estimates use ratio-adjusted projections from partially coded FARS data and may change as final 2025 FARS data is completed. Rural interstate fatality counts are small enough that a 1% change (17 additional deaths) is within the margin of the estimation method. See methodology for caveats.