Waymo's School-Bus Probe Lists 27 Reports in One Case, 8 in Another—They're Not a Trend

Federal auto-safety investigators have opened a new review of Waymo's 2025 school-bus software recall remedy. But the original case records carry an important numerical trap: the preliminary evaluation closed on November 6 lists 27 entries in its “Other” incident-report column, while the recall query opened that day lists 8. Those figures are from different agency case summaries. They do not establish a before-and-after change in school-bus incidents, or show whether the software fix succeeded or failed.¹²

The National Highway Traffic Safety Administration (NHTSA) says a majority of the incidents identified in its earlier preliminary evaluation occurred after November 17, 2025, the date Waymo reported having finished deploying its remedy. That is why the agency opened recall query RQ26006 on November 6, 2026, covering 2,895 Waymo fifth-generation automated-driving systems. Its stated task is to assess the adequacy of the prior recall remedy and the causes of the post-remedy reports—not to announce another recall or a final finding about why those events happened.¹²

Reuters first reported the new inquiry on November 9. The separate numerical comparison below is drawn directly from Waymo's earlier NHTSA-filed recall report and both original agency investigation summaries, rather than treating the headline investigation population as an incident denominator.³⁴¹²

Three documents, three different count scopes

Public record	Reported population	Incident or citation figure actually described
Waymo's December 10, 2025 safety recall filing, 25E084	3,067 potentially involved fifth-generation automated-driving software units	12 citations then known to Waymo, occurring September 25–November 14, 2025. These are the company's historical figures, not a complete later incident tally.4
NHTSA preliminary evaluation PE25013, closed November 6, 2026	2,895 in the closing résumé	27 in the agency's “Other” column, described as an Atlanta media report and citations supplied by Austin Independent School District. Manufacturer and total case-count cells are marked confidential.2
NHTSA recall query RQ26006, opened November 6, 2026	2,895 in the opening résumé	8 in the “Other” column, described as Austin ISD traffic citations. Manufacturer and total case-count cells are also marked confidential.1

The 27 and 8 are not identified as two nonoverlapping sets of events. They are not labeled as all incidents before and after the software update, either. Subtracting 8 from 27 to claim 19 fewer incidents—or adding them to claim 35 distinct cases—would assume a common reporting period, complete totals and known overlap that the two records do not provide.¹²

The 3,067-to-2,895 difference is equally unsuitable as a measure of vehicles repaired, removed or still exhibiting the issue. The first number comes from Waymo's December 2025 recall filing for potentially affected ADS software; the latter is the regulator's population for a distinct November 2026 investigation. Neither filing offers a vehicle-by-vehicle reconciliation of the difference.⁴¹

One further limitation matters in a safety story: the agency documents leave crash, injury and fatality rows without numerical entries while masking certain other counts as confidential. An empty cell is not an independently established zero. Waymo's old filing stated that it knew of no collisions associated with the behavior at that time; that company-submitted historical position does not answer what happened afterward.⁴¹²

The remedy question remains open

Waymo reported to NHTSA in December 2025 that a software update had been deployed between November 5 and November 17, after its system had sometimes resumed moving before a stopped school bus had retracted its stop arm or stopped flashing red lights. The company described an interaction in its driving software between yielding to the bus and avoiding impeding another priority vehicle. That is Waymo's explanation of the earlier identified behavior, not an independently verified explanation for every later report.⁴

NHTSA's new opening document says most of the incidents it identified during PE25013 occurred after the company-reported completion date. Its decision to open an adequacy inquiry is therefore consequential even though the aggregate public incident categories cannot establish a trend or an actual failure rate.¹²

Waymo told TechCrunch, in a statement originally obtained by that publication on November 9, that it had made significant improvements to behavior around school buses and received no citations since March 2026. Waymo also said it would respond to NHTSA's query. The citation statement is the company's position—not an independently audited count of all encounters, reports or potentially unsafe behavior after March.⁵

For passengers and communities, the current verifiable status is a federal investigation of remedy adequacy. The NHTSA documents reviewed here do not announce a second remedy, determine a post-update incident rate or resolve the underlying software-effectiveness question. Those conclusions require additional regulator findings and comparably scoped incident evidence—not arithmetic across unrelated table entries.¹²

Sources

Footnotes

1. NHTSA Office of Defects Investigation, RQ26006 Open Résumé, pages 1–2, dated November 6, 2026. Agency's opened recall query, 2,895 population, 8 “Other” entries, confidential totals, and adequacy-inquiry rationale; not a final defect finding. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹

2. NHTSA Office of Defects Investigation, PE25013 Close Résumé, pages 1–2, dated November 6, 2026. Agency's closed preliminary evaluation, 27 “Other” entries and explanation of transition to RQ26006; not a pre/post-remedy rate series. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸

3. Reuters, U.S. opens probe into 2,895 Waymo vehicles over school-bus safety violations, November 9, 2026. Originating public reporting on the new federal inquiry; agency figures checked against originals below. ↩

4. Waymo LLC / NHTSA, Part 573 Safety Recall Report 25E084, filed December 10, 2025. Company-submitted historical recall population, claimed software rollout, contemporaneous citations, and technical explanation. Not independent verification of later outcomes. ↩ ↩² ↩³ ↩⁴ ↩⁵

5. Kirsten Korosec, TechCrunch, Waymo robotaxis kept illegally passing school buses after software fix, feds say, November 9, 2026, updated with original Waymo spokesperson statement. The company's citation claim has not been independently audited by At The Charger. ↩