Waymo's Flood Problem: A Software Patch That Didn't Hold Water
On May 21, Waymo suspended robotaxi operations across five US cities after a software patch failed to prevent another autonomous vehicle from driving into standing water. The patch, pushed to all 3,791 vehicles in Waymo's fleet less than two weeks earlier, was supposed to be the interim fix for a known flaw. It wasn't enough.
An unoccupied Waymo robotaxi got stuck on a flooded street in Midtown Atlanta during severe storms on May 21, repeating the same failure mode that triggered a recall on May 8 and a service shutdown in San Antonio a month earlier. The company has now paused operations in Atlanta, Austin, Dallas, Houston, and San Antonio (the latter has been offline since late April). It also suspended all freeway rides in San Francisco, Los Angeles, Phoenix, and Miami while it addresses vehicle performance in construction zones.
The Root Cause: No Hard-Stop for Water
According to a letter posted on the National Highway Traffic Safety Administration's (NHTSA) website, the software flaw allows vehicles "to slow and then drive into standing water on higher speed roadways." In the San Antonio incident on April 20, an empty Waymo robotaxi encountered a flooded section of a 40-mph road, detected the water, reduced speed, and then drove into it anyway. Its decision system had no hard-stop condition for water in its path. The vehicle was swept into a creek.
Waymo filed a voluntary recall covering 3,791 robotaxis using its fifth and sixth-generation automated driving systems. The interim software update placed restrictions on operations during elevated flood risk. That update was not enough to prevent the Atlanta incident.
No Permanent Fix Yet
Waymo has admitted it still has no permanent fix. When it filed the recall, the company acknowledged that the "final remedy" for avoiding flooded areas had not been developed. The Atlanta storm produced flooding before the National Weather Service had issued a flash flood warning, watch, or advisory, meaning Waymo's weather-monitoring system (which relies in part on official alerts) had no signal to act on. The company told the BBC that safety was its "highest priority" and that it was "closely monitoring forecasts, alerts, and live weather conditions."
A Pattern of Recalls and Investigations
The flood problem is Waymo's third recall since February 2024. The first covered 444 vehicles after two robotaxis in Phoenix separately struck the same improperly towed vehicle. The second, filed in May 2025, covered 1,212 vehicles involved in low-speed collisions with stationary barriers including parking gates and telephone poles. Two active NHTSA investigations into separate failure modes are ongoing, including one related to a January incident in which a Waymo robotaxi struck a child.
The freeway suspension was triggered by a separate but related concern. A Waymo rider posted on X on May 19 that a robotaxi "freaked out and sped up to highway speeds through construction trucks" on a closed freeway section, with police chasing the vehicle. The rider wrote: "Genuinely thought we were about to die." Waymo has not commented on the specific incident but said it was evaluating its vehicles' freeway performance in construction zones.
Architectural Limitations of Rule-Based Systems
Waymo's approach relies on detailed maps and hand-coded rules for every scenario. The flood incidents illustrate the limitations: a system that maps every road and codes rules for every scenario cannot code for every scenario. Water on a road in a city it has mapped for years exposed a gap that should have been closed long ago.
In contrast, Wayve, the London-based autonomous driving startup backed by Uber, raised $1.5 billion at an $8.6 billion valuation in February and is planning robotaxi pilots in London and Tokyo in 2026. Its AI-first approach learns from driving data rather than relying on detailed maps and hand-coded rules, representing a fundamentally different philosophy.
Industry-Wide Pattern of Failures
Over the past year, driverless car services have faced a pattern of failures that individually look minor but collectively undermine public confidence. In December 2025, a large power outage in San Francisco left Waymo vehicles stalled across the city, blocking intersections. In April, a mass Apollo Go robotaxi outage in the Chinese city of Wuhan caused more than a hundred self-driving cars to stop mid-traffic. Each incident reinforces the same lesson: autonomous vehicles work well within the conditions they were designed for and fail conspicuously when they encounter conditions they were not.
The Credibility Gap
The autonomous vehicle industry is investing billions on the assumption that the technology will eventually handle every scenario a human driver can. Waymo is closer to that goal than any other commercial operator, and its safety record over hundreds of millions of miles is strong by statistical measures. But floods are not edge cases. They are weather. A robotaxi service that cannot operate when it rains heavily in Atlanta or San Antonio, cities where heavy rain is a routine occurrence, is a service that has not yet earned the trust its expansion plans require. The permanent fix, whenever it arrives, will need to solve not just the software problem but the credibility gap that three recalls and five city shutdowns have created.
What's Next for Waymo
Waymo expects to resume freeway routes soon but has not provided a timeline. The company provides more than 500,000 paid trips per week across multiple US cities and plans to expand to San Diego, Las Vegas, and Detroit in 2026, with a goal of offering one million paid rides per week by the end of the year. It also hopes to launch a robotaxi service in London later this year. Uber signed a deal worth up to $1.25 billion for up to 50,000 autonomous Rivian R2 robotaxis. But until the flood problem is fixed, those expansion plans remain on shaky ground.
