Waymo Deploys Fleet-Wide Software Update After Texas Weather Incident

Waymo is updating the software on its entire US fleet of nearly 3,800 robotaxis after an unoccupied vehicle drove into a flooded Texas road in April. Instead of a traditional recall requiring a trip to the mechanic, this fix is being deployed over the air, instantly teaching the entire network how to handle severe flash flooding.[1][2]

For the everyday driver and prospective car buyer, this incident highlights a fundamental shift in what automotive maintenance is about to look like. We are moving into an era where cars learn from a single localized mistake and update overnight. The days of discovering a defect, receiving a notice in the mail, and waiting months for a physical dealership fix are being replaced by digital patches that arrive while the vehicle is parked in the driveway.[1]

The catalyst for this fleet-wide update occurred on April 20 in San Antonio, Texas. During a severe spring storm, an empty Waymo robotaxi encountered a flooded section of road near Salado Creek. While the vehicle's sensors successfully detected the hazard and reduced its speed, the software ultimately failed to halt and reroute the car.[2][4]

The robotaxi proceeded into the water and was swept into the creek, requiring recovery days later. Because the vehicle was in transit without passengers, no injuries were reported. However, the failure of the system to recognize the depth and danger of the untraversable lane prompted immediate internal review at Waymo.[2][3]

Ten days after the incident, Waymo voluntarily filed a recall with the National Highway Traffic Safety Administration (NHTSA). The filing, which became public in mid-May, covers 3,791 vehicles equipped with the company’s fifth- and sixth-generation automated driving systems. The core issue identified was that on higher-speed roadways, the vehicles might slow down but not completely stop when facing flooded lanes.[3][4][6]

What makes this recall notable for the future of car ownership is the nature of the remedy. Waymo’s fix is entirely digital. The company is deploying an over-the-air (OTA) software update that refines the fleet's extreme weather operations, specifically teaching the autonomous system how to better evaluate and avoid standing water.[1][5]

Until the final software patch is fully rolled out, Waymo has implemented strict interim constraints. The company adjusted its vehicle maps and limited access to areas prone to flash flooding. Waymo also temporarily paused its San Antonio operations to review its flood monitoring procedures and safety protocols, ensuring the updated safeguards were robust before resuming service.[1][3][5]

The NHTSA filing also inadvertently provided a rare look at the sheer scale of Waymo's current operations, a metric that matters to urban planners and ride-hail users alike. The document revealed an active fleet of 3,791 robotaxis—nearly double the 2,000 vehicles the Alphabet-owned company reported in late 2025. This expanding footprint now delivers over 500,000 paid rides weekly across ten US cities.[1]

That rapid growth underscores why fleet-wide learning is critical. As autonomous vehicles move from predictable, dry environments like Phoenix into regions with volatile weather patterns like Texas, they will inevitably encounter new edge cases. The San Antonio flood serves as a real-world stress test for how the industry handles these unpredictable scenarios.[4]

For local communities and riders relying on these services, the event demonstrates that while self-driving technology is not flawless, its mitigation process is highly systematized. When one vehicle encounters an unexpected variable, the data is captured, analyzed, and used to immunize the rest of the fleet against the same error.[1][5]

As the broader automotive industry shifts toward software-defined vehicles, this rapid cycle of encounter, recall, and over-the-air resolution offers a preview of the future for all car owners. It illustrates a model where vehicle safety continuously improves over the lifespan of the car, turning isolated incidents into collective upgrades without ever visiting a service center.[1]