NHTSA probes Waymo over school-bus conduct after November 2025 recall

NHTSA’s Office of Defects Investigation (ODI) opened a recall-effectiveness inquiry on Friday into Waymo’s November 2025 software update, after finding that “a majority of incidents occurred after the ... remedy date.”

The chronology turns a completed fleetwide remedy into a second federal test. NHTSA’s newest investigation will determine whether the update fixed the identified behavior and will “identify the causes of the conditions leading to the post-remedy reports.”

The report dates the inquiry only as Friday and says it began several months after Waymo issued the voluntary software recall. It provides no calendar date for Friday, fleet size, or incident-count breakdown. “Post-remedy reports” here means incidents recorded after NHTSA’s remedy date for the update.

The policy stakes are narrow but consequential: regulators must determine whether a software correction eliminated the reported conduct after reaching every vehicle, rather than treating fleet installation as proof of behavioral change.

What started the federal inquiry?

NHTSA’s first probe examined the root issue behind Waymo robotaxis driving past stopped school buses with flashing red lights and deployed stop-sign arms. The agency documented that conduct in Atlanta, Georgia, and Austin, Texas.

The subsequent Waymo recall covered two behaviors: running red lights and proceeding past stop signs from stopped school buses. The original account describes both visible school-bus cues as part of the documented encounters.

That record supplied the baseline for the new inquiry. If incidents continued after the update reached the fleet, NHTSA needed to test both the remedy’s performance and the conditions behind later reports.

How broad was Waymo’s fix?

Waymo voluntarily issued the software recall in 2025. The company installed the update across its entire robotaxi fleet, and the work was completed in November 2025.

The latest ODI investigation references the late-2025 recall, which followed NHTSA’s earlier root-cause investigation. The sequence contains three stages: documented bus encounters, a fleetwide software remedy, and a new inquiry into post-remedy reports.

The cited account does not identify the software version, release name, number of vehicles, or technical mechanism used to change robotaxi behavior. It establishes deployment scope and timing, but not how the software works.

That distinction constrains what an outcome can establish. NHTSA can document that an update reached the whole fleet without yet resolving why reports continued or whether another remedy is required.

Why did NHTSA revisit the completed recall?

The recall query began on Friday for one stated reason: “a majority of incidents occurred after the ... remedy date.” The phrase makes the timeline the agency’s core evidence for revisiting a completed recall.

The new investigation has two stated objectives. First, it will determine whether Waymo’s update actually fixed the problem. Second, it will identify the causes of the conditions that produced reports after the remedy.

The two objectives separate outcome from cause. A completed installation does not, by itself, explain later incidents, and the source records no preliminary finding on either effectiveness or causation.

The opening does not establish that the recall failed, nor does the source report announce a new defect finding or enforcement action against Waymo. It records an open investigation into a chronology the agency considers unresolved.

Where did the school-bus incidents occur?

NHTSA’s first findings place the documented encounters in two cities, with flashing red lights and deployed stop-sign arms:

• Atlanta, Georgia: One of two locations where NHTSA documented a Waymo robotaxi driving past a stopped school bus.
• Austin, Texas: The second city included in the agency’s findings involving the same stopped-school-bus conditions.

The original account provides no number of encounters, dates for individual events, injury outcomes, or total fleet figures for either city. Without those figures, the reported pattern cannot be converted into an incident rate or compared across the two locations.

What does the Santa Monica collision probe cover?

Separately, NHTSA’s Office of Defects Investigation opened another probe into an incident in which a Waymo robotaxi hit a child in Santa Monica, California, in January.

The source gives January but no year and does not describe the vehicle, the child’s injuries, the sequence of events, or any findings. Those details remain unspecified in the account presented here.

The collision case remains distinct from the school-bus review and the recall-effectiveness inquiry. The source does not link the January incident to a stopped school bus or to the software update.

Placing the cases together does not merge their evidence. It does show that NHTSA is examining two consequences attributed to Waymo robotaxis: interaction with school-bus stop signals and a collision involving a child, each through a separate probe.

What is the NTSB examining?

The National Transportation Safety Board (NTSB), described in the source as an independent federal agency, launched its own investigation earlier this year into Waymo robotaxi behavior around stopped school buses.

That creates parallel federal scrutiny. NHTSA is testing the fleet update and a separate child-collision incident, while the NTSB is examining stopped-bus behavior described in NHTSA’s earlier findings.

The source does not connect the NTSB inquiry to the recall-effectiveness query, identify shared investigators, or provide a timetable for either case. It likewise gives no count of NTSB incidents.

Two agencies can examine the same operating condition without issuing the same finding. Here, the public account leaves their evidence, methods, and conclusions separate.

Where does Waymo stand?

The source report states that it asked Waymo for comment and will update the story when the company responds. No Waymo quote, technical explanation, acknowledgment of the child collision, or response to the recall inquiry appears in the source.

Waymo’s response could address whether its fleet update targeted the documented signals, but NHTSA has not recorded that response in the material cited here. The investigation proceeds from the agency’s post-remedy chronology, not from a public company admission.

Any later Waymo statement would add context rather than supply the missing causal finding. NHTSA still has to determine what produced the post-remedy reports, even with a completed recall and full-fleet installation.

What are the oversight stakes?

For autonomous-vehicle oversight, the case illustrates a demanding evidentiary standard for software remediation. Federal investigators can revisit a voluntary fleetwide update when later reports suggest the original problem persisted. The source provides no comparable cases, so it does not establish a broader enforcement trend.

For Waymo, the unresolved questions are more immediate. The company completed an update across its entire fleet in November 2025, yet NHTSA says a majority of incidents followed the remedy date. Fleet coverage and incident reduction are not the same measure.

The case also carries a direct physical-safety dimension. NHTSA’s separate Santa Monica inquiry concerns a robotaxi striking a child, while the NTSB inquiry centers on behavior near stopped school buses. Together, they put collision risk and school-bus interactions alongside the software-effectiveness question.

What comes next?

NHTSA’s findings on the November 2025 remedy, any Waymo technical response, and the outcomes of the Santa Monica and NTSB investigations will provide the next measurable evidence on those safety questions.

The immediate test is narrower than a judgment on autonomous driving overall. It asks whether a specific November 2025 software update resolved a specific class of school-bus encounters, and why NHTSA continued to receive reports afterward.