Waymo is recalling nearly 3,800 autonomous taxis in the United States after identifying a safety issue involving flooded roads during severe weather conditions. The issue came to light after one of Waymo’s autonomous vehicles reportedly drove into a flooded lane in San Antonio, Texas, during intense rainfall in April. According to the company, the vehicle was unoccupied at the time, and no injuries occurred. Even without injuries, the incident immediately raised concerns about how autonomous systems interpret dangerous road conditions during extreme weather. Flooding remains one of the most unpredictable hazards for both human drivers and self-driving technology, especially when water depth and road visibility rapidly change. Waymo says it has already implemented temporary safeguards while developing a permanent software fix. Regulators are also closely monitoring the situation as scrutiny surrounding autonomous vehicle safety continues to increase across the industry. Flooded Roads Exposed A Weakness In The System According to National Highway Traffic Safety Administration filings, the recall specifically targets scenarios where Waymo vehicles could potentially enter flooded roads with higher speed limits. The April incident in San Antonio prompted the company to review how its software handles similar situations involving standing water, fast-moving traffic, and limited route alternatives. Severe weather can create conditions where lane markings, curbs, and road edges become difficult for sensors and cameras to interpret accurately. Waymo stated that it is now refining its “extreme weather operations” while limiting vehicle access to areas where flash flooding may occur. The company has also updated portions of its operating maps and temporarily reduced the geographic scope of service in certain conditions. Unlike traditional recalls involving physical defects, this recall focuses entirely on software behavior and operational safety protocols. That has become increasingly common as modern autonomous vehicles rely heavily on constantly evolving software systems. Autonomous Driving Still Struggles