Incident Overview
On April 20, 2026, an unoccupied Waymo robotaxi entered a flooded roadway near Salado Creek in San Antonio, Texas, during severe storm conditions. The vehicle was subsequently swept into a nearby waterway after detecting but proceeding through floodwater at a reduced speed. No passengers were aboard the vehicle at the time, and no injuries were reported. The incident occurred during a period of extreme weather, highlighting the challenges autonomous vehicles face in accurately perceiving and responding to environmental hazards beyond typical operational parameters.
The vehicle was equipped with Waymo’s fifth- or sixth-generation automated driving system, which relies on a combination of sensors, cameras, and AI-driven perception algorithms to navigate road conditions. Despite detecting the floodwater as potentially untraversable, the system allowed the vehicle to continue forward at a reduced speed before it was carried into the waterway.
Waymo’s response to the incident included an immediate operational pause in San Antonio and a broader recall of 3,791 vehicles equipped with its fifth- and sixth-generation automated driving systems. The recall was filed with the National Highway Traffic Safety Administration (NHTSA) and cited a software issue that could permit vehicles to drive onto flooded roads. This recall marked one of the largest fleet-wide actions taken by Waymo in response to a single incident, signaling the severity of the perceived risk.
The incident became widely discussed within the autonomous vehicle industry, as it represented one of the most-shared robotaxi failures of 2026. It also prompted Waymo to temporarily suspend operations in five additional U.S. cities, where similar risks of flooding or adverse weather conditions were identified. The company’s actions reflected a cautious approach to mitigating potential harm to passengers, pedestrians, and property in the event of repeated failures under extreme conditions.
Regulatory and industry observers noted that the incident exposed limitations in current autonomous vehicle systems when confronted with edge cases involving natural disasters. The OECD AI Incidents and Hazards Monitor (AIM) classified the event as an AI-related hazard, emphasizing the role of Waymo’s automated driving system in the incident. The monitor categorized the implicated systems as the Waymo Driver, robotaxi systems, and fifth- and sixth-generation automated driving systems, all of which were developed by Waymo and deployed in operational environments.
The floodwater event also raised questions about the adequacy of existing safety protocols for autonomous vehicles in unpredictable weather scenarios.
Possible Root Cause
The primary root cause of the incident appears to lie in the automated driving system’s handling of floodwater detection and response. According to Waymo’s statements to regulators, the vehicle’s sensors identified the floodwater as potentially untraversable but proceeded at a reduced speed rather than initiating an immediate stop or alternative route. This suggests a gap in the system’s ability to assess risk accurately when confronted with rapidly changing or extreme conditions.
However, the floodwater’s depth and flow rate exceeded the system’s ability to predict or mitigate the risk of being swept into the waterway. This indicates a failure in the system’s dynamic risk assessment, where the perceived threat of floodwater did not trigger a sufficiently aggressive avoidance response.
Another contributing factor may be the limitations of sensor fusion in extreme weather. In this case, the combination of these factors likely impaired the vehicle’s ability to assess the true severity of the floodwater, leading to an underestimation of the risk posed by the waterway.
Waymo’s recall of 3,791 vehicles equipped with fifth- and sixth-generation automated driving systems suggests that the issue extends beyond a single vehicle or software version. The recall notice filed with NHTSA indicated that the software could allow vehicles to drive onto flooded roads, implying a systemic vulnerability in the floodwater detection and response logic. This systemic nature of the problem suggests that the root cause may involve flaws in the design or implementation of the perception and decision-making algorithms across multiple generations of Waymo’s autonomous driving systems.
While the system may have adhered to predefined safety parameters under normal conditions, the floodwater scenario presented an unprecedented challenge. The lack of a clear protocol for handling such extreme events may have contributed to the vehicle’s inability to prevent its entry into the waterway.
Takeaways
The San Antonio floodwater incident serves as a critical case study in the limitations of current autonomous vehicle technology when faced with natural disasters or extreme environmental conditions. It demonstrates that even advanced AI-driven systems can encounter unforeseen challenges in real-world scenarios, particularly when sensor data is unreliable or risk assessment models are insufficiently comprehensive.
One key takeaway is the necessity for autonomous vehicle developers to incorporate more robust contingency planning for edge cases involving severe weather. This includes enhancing floodwater detection algorithms to better differentiate between traversable and untraversable conditions, as well as implementing fail-safe mechanisms that prioritize immediate avoidance over gradual deceleration when high-risk scenarios are identified.
Waymo’s recall and operational pauses indicate a recognition of the need for continuous monitoring and updates to ensure that vehicles can respond appropriately to evolving environmental threats. This may involve integrating additional sensor modalities or machine learning models trained on rare but high-impact events, such as flash flooding.
The NHTSA’s involvement in the recall process highlights the role of regulatory bodies in holding developers accountable for systemic vulnerabilities.
For Waymo and other autonomous vehicle developers, the incident serves as a reminder of the importance of transparency and proactive risk mitigation. By publicly acknowledging the limitations of their technology and taking decisive action—such as fleet recalls and operational pauses—Waymo demonstrated a commitment to safety that could help rebuild trust with regulators, passengers, and the public.
Lessons learned from this event could lead to improvements in how autonomous systems handle extreme conditions, ultimately enhancing their reliability and reducing the potential for harm to passengers, pedestrians, and property.
Sources
Incident summary from the AI Incident Database, used under CC BY-SA 4.0.