Waymo’s flood recall shows robotaxis still struggle with judgment Subscribe to our free newsletter today to keep up to date with the latest transportation and logistics news. Waymo’s latest recall is not just a software event. It is a reminder that autonomous vehicles are moving from controlled technical achievement into the messier world of public infrastructure, changing weather and split-second risk judgment. The Alphabet-owned robotaxi company recalled 3,791 vehicles after a Waymo vehicle entered a flooded roadway in San Antonio on April 20, according to filings with the National Highway Traffic Safety Administration. The vehicle was unoccupied and no injuries were reported, but the system had detected the flooded road and still continued into the water at a reduced speed. For an industry built on the promise that software can make mobility safer, the incident is awkward. Floodwater is not an obscure hazard. Human drivers are repeatedly warned not to enter it because depth, road surface and current are difficult to judge from the driver’s seat. For an autonomous system, the problem is harder. Cameras, lidar, radar, maps and prediction models can identify a hazard, but the central question is whether the vehicle can interpret the scene conservatively enough when the safest choice is to stop or reroute. The recall is small in market terms, but large in strategic meaning Waymo’s recall covered fifth- and sixth-generation automated driving systems, according to NHTSA documents and subsequent reporting. The company said it added operational restrictions after the April 20 event and its safety board decided on April 24 to conduct a recall. The numbers matter because they reveal the scale of Waymo’s current US operation. A fleet of 3,791 vehicles is still modest compared with conventional ride-hailing networks, but it is large enough that rare failures are no longer hypothetical. When autonomous