NHTSA Closed Two Safety Probes on a Saturday Afternoon. Both Companies Had Already Recalled.

Saturday, June 27, 2026. NHTSA closed two major vehicle safety investigations covering 817,243 vehicles. Nobody held a press conference, and no administrator appeared on camera. Two Reuters wire stories, published into the void of a summer weekend, carried the entire disclosure.

Investigation one: Tesla Model 3 and Model Y power steering. NHTSA opened a preliminary evaluation in July 2023 after owners reported their steering wheels becoming difficult or impossible to turn.^[1] The agency upgraded it to a full engineering analysis in early 2024, covering 376,241 vehicles from the 2023 model year. Tesla recalled those vehicles in early 2025, pushing an over-the-air software update to prevent overvoltage breakdown on the motor drive circuit board. Then Tesla said something remarkable: the recall "was not in response to NHTSA's investigation."^[1]

NHTSA kept the investigation open for another sixteen months after the recall, then closed it Saturday.

Then there was the Honda Odyssey side airbag probe. NHTSA opened a probe in October 2025 after complaints that side airbags were detonating unexpectedly while the minivan was in motion, sometimes triggered by potholes.^[2] That probe covered 441,002 Odysseys from the 2018 through 2022 model years. Honda recalled in April 2026, and by that point, Honda had logged 130 warranty claims, 25 injuries, and zero fatalities going back to January 2017.^[2] NHTSA closed its evaluation. Also Saturday.

The Pattern

Since the National Traffic and Motor Vehicle Safety Act passed in 1966, NHTSA has participated in recalls affecting more than 390 million vehicles.^[3] The language is always "participated in," because the agency almost never forces a recall through the courts, its statutory authority under 49 USC §30118 notwithstanding. The choreography is consistent: NHTSA opens investigation, manufacturer issues voluntary recall, NHTSA closes investigation. Both parties declare victory: manufacturers avoid the stigma of a government-ordered recall, and NHTSA avoids the legal cost and risk of a court fight. Consumers get a fix, eventually, if they bring their car to a dealer, which roughly 25% never do.^[4]

Tesla's explicit declaration that its recall was "not in response to" the investigation is the system's subtext becoming text. It amounts to saying: we would have done this anyway, you didn't make us, and the thirty-five months you spent investigating were largely irrelevant to the outcome.

The Exposure Window

While NHTSA investigated, people drove. A calculation of the exposure:

Tesla's 376,241 affected vehicles spent approximately eighteen months on roads with a known steering defect before the recall shipped in early 2025. That produces roughly 564,000 vehicle-years of exposure at an average of 11,500 annual miles per vehicle, or about 6.5 billion vehicle miles traveled during the pre-recall window alone.^[5]

Honda's 441,002 Odysseys carried the side-airbag defect from the time the complaints surfaced until the April 2026 recall. NHTSA's own probe ran eight months, but Honda's warranty data shows the problem existed since January 2017, which means those minivans spent up to nine years on the road with airbags that could detonate over a pothole.

Apply the 2024 national fatality rate of 1.19 deaths per 100 million VMT.^[5] During just the eighteen-month Tesla pre-recall window, roughly 77 people died in crashes involving those 376,241 vehicles. Not 77 people killed by the steering defect. Seventy-seven people who died in crashes where the affected vehicles were present, while the investigation inched forward. The defect-attributable fraction is unknown and likely small. But "likely small" is not "zero," and the exposure window is not hypothetical.

What's Still Open

Two probes closed, three remain active for Tesla alone. Investigation EA26002, covering 3.2 million vehicles for FSD camera failures in sun glare and fog, sits at the engineering analysis stage, the final step before NHTSA can demand a recall.^[6] A separate preliminary evaluation, PE25012, covers 2.88 million vehicles for FSD-related traffic violations. A third probe just opened after a fatal crash in Katy, Texas, where a Model 3 driver claimed FSD drove the car into a house.^[6]

Combined scope of active Tesla investigations: approximately 6.1 million vehicles.

Limitations

The exposure-window fatality estimate uses national average rates, not model-specific rates. Tesla Model 3 and 2018-2022 Honda Odyssey are not present in our FARS 2014-2023 per-model dataset as distinct cohorts, so we cannot calculate vehicle-specific fatality rates for the affected populations. The 77-death estimate for the Tesla window represents total traffic fatalities involving these vehicles, not deaths attributable to steering loss. The actual defect-caused fraction could be one death, or none, or something NHTSA will never quantify because the investigation is now closed.

Strongest Counterargument

NHTSA's investigation pressure is what causes voluntary recalls in the first place. Without the agency opening a preliminary evaluation, Tesla may never have recalled. The fact that manufacturers recall "voluntarily" is evidence the system works: the credible threat of enforcement produces compliance without the cost of litigation, and consumers get fixes faster than they would through court-ordered recalls. Closing investigations on a Saturday is routine administrative housekeeping, not a calculated news dump.

That argument holds if you believe NHTSA's investigation was the causal factor. Tesla explicitly said otherwise. And twenty-five percent of recalled vehicles never get fixed regardless of which day the press release drops.^[4]

What You Should Do

If you own a 2023 Tesla Model 3 or Model Y, check that your vehicle has received the steering-related OTA update. In the vehicle's software menu, confirm you're running the latest firmware. If you own a 2018-2022 Honda Odyssey, check your VIN at nhtsa.gov/recalls and schedule the dealer visit. Side airbags detonating over potholes is not a hypothetical risk. Twenty-five people were injured.