NHTSA Opens Recall Query Into 2,895 Waymo Robotaxis Over Continued School Bus Safety Violations

Federal safety regulators have launched a formal investigation into 2,895 Waymo robotaxis following reports that the autonomous vehicles are still passing stopped school buses. The National Highway Traffic Safety Administration opened the recall query on Monday to determine why a previous software fix failed to resolve the safety violations.[1][4]

The inquiry centers on a fleet of Jaguar I-Pace vehicles operating across multiple major American cities. Waymo previously issued a voluntary software recall in 2025 to address the exact same issue, deploying an over-the-air update intended to improve the system's recognition of deployed school bus stop arms.[2][5]

Despite that digital intervention, the federal agency received new incident reports indicating the autonomous driving system continues to misinterpret or ignore the flashing red lights and extended stop signs of active school buses. The ongoing violations prompted the Office of Defects Investigation to escalate its oversight of the Alphabet-owned company.[3][6]

"The Office of Defects Investigation has opened a recall query to evaluate the performance of the remedy," the NHTSA stated in its public filing. The agency noted that the investigation will assess whether the 2025 software update was sufficient to ensure compliance with state traffic laws regarding school buses.[4][7]

The Edge Case of School Buses

School buses present a uniquely complex challenge for autonomous vehicle perception systems. Unlike static traffic signals or standard stop signs, a school bus dynamically deploys its stop arm and activates flashing lights, creating a temporary, high-stakes exclusion zone that human drivers are trained to recognize instantly.[1][5]

Waymo's sensor suite relies on a combination of lidar, radar, and high-resolution cameras to build a three-dimensional map of its surroundings. While these systems excel at tracking vehicle trajectories and pedestrian movement, interpreting the specific sequence of a school bus halting and deploying its signals requires specialized machine learning models.[2][6]

The 2025 software update was specifically engineered to refine how the automated driving system processes these visual cues. Engineers adjusted the software weights to prioritize the detection of the yellow school bus profile and the subsequent deployment of the red octagonal stop arm.[3][4]

However, real-world variables such as glare, partial occlusion by other vehicles, and varying state-by-state designs for school bus lighting can complicate detection. The NHTSA probe suggests that the updated algorithms are still encountering edge cases where the vehicle fails to initiate the required braking sequence in time.[5][7]

Expanding Fleets and Federal Scrutiny

The timing of the federal query coincides with Waymo's aggressive expansion of its commercial operations throughout 2026. The company currently operates fully driverless ride-hailing services in San Francisco, Phoenix, Los Angeles, and Austin, logging millions of autonomous miles each month without a human safety operator behind the wheel.[1][3]

As the fleet footprint grows, so does the statistical likelihood of encountering active school bus routes during morning and afternoon transit hours. City officials in these operational zones have closely monitored the robotaxis, balancing the promise of reduced traffic fatalities against the risks of deploying beta software on public roads.[2][6]

Waymo has consistently defended its safety record, pointing to data that shows its vehicles are involved in fewer injury-causing collisions than human drivers over comparable distances. A company spokesperson addressed the new federal query on Monday, emphasizing their cooperative stance with regulators.[4][5]

"We are proud of our safety record and will continue to work closely with NHTSA on this inquiry," the Waymo representative stated. The company maintains that its continuous deployment model allows it to rapidly iterate and push safety improvements to the entire fleet simultaneously.[1][7]

The Limits of Over-the-Air Fixes

The current investigation highlights a growing regulatory debate over the efficacy of over-the-air software updates as a mechanism for automotive recalls. Historically, a vehicle recall required a physical trip to a dealership to replace a defective mechanical component, ensuring the fix was definitively applied and tested.[2][4]

In the era of software-defined vehicles, companies like Waymo and Tesla frequently resolve safety defects by transmitting new code directly to the vehicle's onboard computers. While this approach is highly efficient, the NHTSA is increasingly questioning whether these digital patches adequately address the root causes of complex behavioral failures.[3][6]

If the Office of Defects Investigation determines that the 2025 software update was inadequate, Waymo could be forced to implement more restrictive operational boundaries. This might include geofencing the vehicles away from known school zones during specific hours or requiring a more fundamental overhaul of the perception stack.[5][7]

The agency has the authority to demand a second, more comprehensive recall if the data proves the initial remedy fell short. Such an outcome would represent a significant regulatory setback for Waymo, which has largely avoided the severe federal crackdowns that sidelined competitors like General Motors' Cruise.[1][4]

Industry-Wide Implications

The broader autonomous vehicle industry is watching the NHTSA's handling of the Waymo case closely. The standard established here will likely dictate how federal regulators evaluate the safety claims of other robotaxi operators attempting to scale their services across the United States.[2][5]

Competitors such as Zoox and emerging autonomous trucking firms face the exact same perception challenges regarding school buses and emergency vehicles. A ruling that software patches are insufficient for certain critical safety failures could force the entire sector to rethink its validation and testing protocols.[3][7]

AI

Safety advocates argue that the tolerance for error around school buses must remain absolute zero. The unpredictability of children boarding and exiting the vehicles creates an environment where a single perception failure by a 4,800-pound autonomous vehicle could result in a catastrophic outcome.[4][6]

"Passing a stopped school bus is one of the most dangerous traffic violations any driver can commit," noted a safety analyst reviewing the federal filing. The analyst emphasized that autonomous systems must demonstrate superhuman reliability in these specific scenarios before achieving widespread public trust.[1][5]

Data Collection and Next Steps

Over the coming months, the NHTSA will compel Waymo to hand over extensive telemetry data, internal testing logs, and video recordings from the 2,895 vehicles under investigation. Investigators will analyze the exact sensor inputs and computational decisions made by the vehicles in the moments before they bypassed the stopped buses.[2][4]

The agency will also review the specific engineering changes Waymo implemented during the 2025 recall. Regulators want to understand why the company's internal simulation and closed-course testing failed to predict that the software update would not fully resolve the issue on public streets.[3][6]

Waymo faces a strict timeline to respond to the federal information requests. Failure to provide comprehensive data or attempting to obscure the frequency of the violations could result in substantial civil penalties and further restrictions on their commercial operating permits.[5][7]

The resolution of this recall query will serve as a critical checkpoint for the commercialization of driverless technology. As the NHTSA transitions from monitoring pilot programs to regulating mature commercial fleets, the agency is signaling that it will hold autonomous systems to an increasingly rigorous standard of compliance.[1][4]

Navigating State Traffic Laws

The complexity of the investigation is compounded by the fragmented nature of traffic regulations across Waymo's operational territories. While the fundamental rule of stopping for a school bus is universal, the specific triggers—such as the distance required to stop or the rules for divided highways—vary significantly between California, Arizona, and Texas.[2][5]

Waymo's automated driving system must maintain a localized understanding of these distinct legal frameworks while navigating in real-time. When a robotaxi crosses from a standard two-lane road onto a multi-lane boulevard with a physical median, the software must instantly recalculate whether a stop is legally required for a bus on the opposite side.[3][6]

Federal investigators will likely examine whether the continued violations stem from a failure to detect the bus entirely, or a failure to correctly apply the local jurisdiction's specific traffic laws once the bus is identified. The distinction is crucial for determining how the software must be rewritten.[4][7]

The outcome of this technical audit will shape the next generation of autonomous vehicle software architecture. Until the NHTSA concludes its review, Waymo's fleet will continue to operate under heightened federal scrutiny, carrying passengers while regulators parse the code that governs their every move.[1][5]