Waymo Returns to San Antonio After Flood Shutdown and Software Fix
Waymo has restarted its robotaxi service in San Antonio, Texas, five months after flash flooding swept an unoccupied vehicle away and forced its longest local service pause to date. TechCrunch
The resumed service covers a 60-mile area that includes downtown and San Antonio International Airport. It is operating with dozens of vehicles and is initially available to some riders. Waymo is now offering paid rides in 15 U.S. cities.
The shutdown traces to heavy rains on April 20. Multiple Waymo robotaxis got stuck during flooding in San Antonio. One unoccupied vehicle was washed into Salado Creek. Houston Public Media The vehicles slowed as they approached waterlogged streets but did not stop.
Waymo briefly paused operations in other Texas cities around the same period and has since resumed those services. It said in late April that it was reviewing its San Antonio operations to refine flood monitoring mitigations. GovTech
The engineering response followed in May. Waymo issued a recall covering almost 3,800 robotaxis after the San Antonio incident. Houston Public Media It made software changes to help its vehicles better detect and avoid standing water and restricted its robotaxis from driving in areas with elevated risk of flooding.
The restart arrives alongside updated safety figures from Waymo. Its analysis covers more than 220 million fully autonomous miles through the end of March 2026 across five operating geographies, with Atlanta included for the first time. The company now drives more than 4 million miles every week. Waymo
Compared to human drivers in the same areas over the same period, Waymo reports 94% fewer crashes causing serious or fatal injuries, 82% fewer crashes in which an airbag deployed, and 82% fewer crashes involving any reported injury. For vulnerable road users, it reports 84% fewer injury-causing crashes involving cyclists, 84% fewer involving motorcyclists, and 93% fewer involving pedestrians.
In Atlanta, the analysis covers more than 5.4 million fully autonomous miles. There, Waymo reports 86% fewer injury-involving crashes and 94% fewer airbag-deployment crashes than the human benchmark.
That operating history frames a period of geographic expansion. In August 2026, Waymo said it would begin testing its autonomous vehicles in Munich in the coming weeks. Reuters It has set a goal to surpass 1 million weekly paid trips by the end of 2026. CBS Austin Driverless passenger service for employees in Austin began in March 2024.
The broader context here is familiar to anyone who has watched automated systems meet uncontrolled environments. Collision avoidance against other road users is a different problem from environmental hazards like standing water. Depth is hard to estimate from cameras and lidar returns off a flat, reflective surface, and safe behavior often requires route-level avoidance rather than a last-meter brake decision.
In this author's view, the San Antonio sequence is instructive precisely because no one was hurt and the failure was operationally legible. Geofencing out of flood-prone segments, tighter integration of weather and hydrology feeds into dispatch, and more conservative thresholds for water depth are incremental fixes. They narrow the operational design domain in the short term to preserve it over the long term. Worth flagging for operators is the pattern: the fleet slowed correctly but lacked a stop and reroute rule with sufficient margin. That is the kind of edge case that only appears at scale, and Waymo's mileage volume now surfaces it faster than road testing alone ever could.
