A flooded road exposed the test Waymo’s robotaxis still have to pass

On April 20, a Waymo robotaxi drove into a flooded section of San Antonio, triggering the company’s first major recall of 3,791 vehicles. The incident, which resulted in no injuries, exposed a fundamental limitation in autonomous driving: the inability to reliably detect and navigate standing water. Waymo’s response—a voluntary recall filed with the National Highway Traffic Safety Administration (NHTSA)—was not a hardware fix but a software patch, deployed over-the-air (OTA) across its entire fleet within days.

No vehicles required physical servicing, a testament to the company’s remote-update infrastructure Electrek.

The recall is less about the incident itself and more about the broader challenge of scaling autonomous vehicles in unpredictable environments. Flooded roads are not a rare occurrence in the U.S., particularly in cities like San Antonio, where flash floods are common. Yet, most self-driving demos take place in dry, well-mapped urban areas. Waymo’s interim constraints—likely limiting operations in heavy rain or flood-prone zones—suggest the company is still refining its environmental sensing capabilities.

The question is whether these fixes are reactive or part of a proactive strategy to handle weather-related edge cases.

The source material also shows that waymo’s OTA recall is a double-edged sword. On one hand, it demonstrates the company’s ability to respond rapidly to safety concerns without grounding its fleet. The fact that 3,791 vehicles can receive a software update simultaneously is a logistical feat, one that traditional automakers would struggle to match.

On the other hand, the incident reveals a critical blind spot in autonomous driving: the reliance on lidar, radar, and cameras to interpret the world. These sensors, while effective in most conditions, can be fooled by reflective surfaces like water, leading to miscalculations in depth perception and path planning.

The San Antonio pause is not just a PR setback but a technical reality check. Waymo has set an ambitious goal of delivering 1 million weekly rides by the end of the year, expanding into new markets like Los Angeles and Austin. Yet, incidents like this underscore the gap between controlled demos and real-world deployment. Flooded roads are not an outlier—they are a predictable challenge in many U.S. cities. If Waymo cannot reliably handle six inches of standing water, its expansion plans may face delays or additional regulatory scrutiny.

The company’s next move—resuming San Antonio service only after reviewing flood monitoring procedures—will be a key indicator of whether it can turn this recall into a lesson rather than a liability.