Waymo deployed its school-bus fix to all affected vehicles. Most identified incidents came later.

Federal auto-safety regulators have opened a recall query into 2,895 Waymo fifth-generation automated-driving systems after finding that most school-bus incidents identified in an earlier investigation occurred after the company said it had finished deploying a software remedy. The November 6 proceeding asks whether that earlier fix was adequate; it is not a finding that the remedy failed.

The distinction matters because Waymo owns the affected fleet. Its December 2025 recall report said a defect could let a robotaxi resume moving before a stopped school bus deactivated its flashing lights or retracted its stop arm. Waymo said a hotfix reached all affected vehicles between November 5 and November 17, 2025.

That report covered 3,067 software units and listed 12 citations dated from September 25 through November 14, 2025, with no related collisions. Because Waymo owned every affected vehicle, the company said conventional owner and dealer notices were unnecessary.

Software delivery was therefore administratively complete. Safety effectiveness was not. NHTSA’s closing report for its preliminary evaluation says a majority of the incidents it identified occurred after November 17, when Waymo reported completing the remedy. The agency closed that evaluation and opened the new recall query to assess the remedy and the cause of the post-remedy reports.

One later incident has a separate federal record. The National Transportation Safety Board says a Waymo vehicle in Austin stopped near a school bus on January 12, 2026, then resumed after a remote-assistance response of “No” to whether the bus signals were active. The vehicle passed while the stop arms remained extended. No crash occurred.

The strongest supported criticism is not that regulators have proved Waymo’s code defective. It is that a company-owned, over-the-air recall can reach 100 percent deployment without proving that the safety behavior has been corrected. When the manufacturer is also the fleet owner, public oversight needs an outcome test—not just a software-distribution receipt.

Waymo’s contrary evidence is material. The company told Reuters and TechCrunch that it has significantly improved how its vehicles behave around school buses, has received no citations since March 2026, will cooperate with NHTSA and believes its automated driver has a positive overall safety impact. Neither the recall report nor the NTSB school-bus record reports a crash or injury.

Important limits remain. NHTSA has opened an inquiry, not ordered a new recall or made a final remedy-failure finding. Some manufacturer incident totals are confidential, so the public record does not support a definitive total. The records also do not explain why the 2025 recall covered 3,067 units while the new query covers 2,895.

The next accountability test is observable: NHTSA should disclose how it will measure remedy effectiveness, what changed after the November 2025 hotfix and why post-remedy reports continued. Waymo should publish version-level performance evidence around stopped school buses. A completed software push is a distribution fact; whether children were protected is the safety result.