Waymo's Rain Problem Is the Whole Robotaxi Industry's Problem

Waymo has paused service in four cities because its robotaxis are struggling with heavy rain and flooded roads, TechCrunch reported on 21 May 2026.

That is a small number of cities and a very ordinary kind of weather. It is also the most useful thing anyone has published about autonomous driving this year. The sequence matters. One of Waymo's robotaxis was spotted driving through a flooded street in Atlanta on the Wednesday before the report, according to local news accounts cited by TechCrunch. It then got stuck for about an hour. The vehicle was recovered and removed. Waymo told TechCrunch it paused service in Atlanta, just as it has in San Antonio, Texas, while it works out a solution. Dallas and Houston were halted because of severe weather across Texas that week, a move first reported by Bloomberg News and later confirmed by the company. A Waymo spokesperson said the Dallas and Houston pauses were made out of caution for the forecast.

A recall with no final remedy

What makes the Atlanta incident worth reading closely is the timeline. Waymo had already issued a software recall for this exact class of failure. And it admitted at the time that it had not finished developing a "final remedy" for avoiding flooded areas, according to documents released by the National Highway Traffic Safety Administration. Instead it shipped an update that placed "restrictions at times and in locations where there is an elevated risk of encountering a flooded, higher-speed roadway." Then a car drove into a flooded intersection anyway.

Waymo's explanation is that the Atlanta storm dumped rain so fast that flooding was already happening before the National Weather Service had issued a flash flood warning, watch or advisory. The company says those alerts are part of a broader set of signals it uses to prepare vehicles for bad weather. That is a reasonable point about the limits of public warning systems. It is also an admission that the vehicle's own perception is not yet the thing standing between the car and deep water.

"Safety is Waymo's top priority, both for our riders and everyone we share the road with. During a period of intense rain yesterday in Atlanta, an unoccupied Waymo vehicle encountered a flooded road and stopped," the company said in a statement.

NHTSA is watching. "NHTSA is aware of this incident, is in communication with Waymo, and will take appropriate action if necessary," a spokesperson for the regulator told TechCrunch.

This is a pattern, not a one-off

The flood pause is not the first time Waymo has struggled to stamp out problematic behaviour quickly. When people noticed Waymo robotaxis illegally passing stopped school buses last year, the company shipped a fix that was supposed to address the issue, and its fleet kept making illegal manoeuvres around school buses. That behaviour sits at the centre of one of two active investigation sets. Both NHTSA and the National Transportation Safety Board are looking into it. Waymo has produced a batch of documents for NHTSA, all redacted to the public. On 15 May, NHTSA sent a second document request because the initial response "necessitates that [NHTSA] receive further data and information."

The second set of investigations concerns a 23 January incident in which a Waymo robotaxi crashed into a child in Santa Monica, California. Waymo has said the vehicle braked to around six miles per hour before striking the child, and that she suffered minor injuries. And the company's own filings show the pattern continuing. Waymo recalled nearly 4,000 robotaxis in June after identifying at least 13 instances of its vehicles driving into highway sections closed for construction. Six happened in Phoenix in April, seven in San Francisco in May. Waymo pulled its robotaxis from all highways on 19 May, and the fix was still "currently under development" in the NHTSA filings. That was the sixth recall Waymo had issued for its robotaxis.

Why the data problem is structural

Set that against the scale of the expansion. Waymo says its vehicles have driven more than 170 million miles autonomously and claims a 13x reduction in serious-injury-or-worse crashes compared with human drivers. It is planning to launch in more than 20 cities this year alone, including London and Tokyo. More cities means more weather, more construction signage, more local road quirks, more of everything that is not in the training set.

This is the part of the story the industry has been slow to say out loud. Self-driving systems are shifting away from rules-based operation and toward reinforcement learning, and as that happens, real-world driving data becomes the scarce input. As TechCrunch reported in January 2026, Uber launched a division called AV Labs to collect exactly that data for its more than 20 autonomous vehicle partners. Uber's chief technology officer, Praveen Neppalli Naga, framed it bluntly: "Our goal, primarily, is to democratize this data, right? I mean, the value of this data and having partners' AV tech advancing is far bigger than the money we can make from this." The division started with one car, a Hyundai Ioniq 5, and engineers still screwing on lidars, radars and cameras. Uber's VP of engineering, Danny Guo, said the team did not know whether the sensor kit would fall off. That is a scrappy operation, and it is also a signal about where the bottleneck sits. Fleet size creates a physical ceiling on how much data a robotaxi company can collect. Simulation helps, but it does not generate the strange scenarios that real roads do.

Regulators are moving too

Waymo's troubles are the most visible, but not the only ones shaping the rules. Tesla is fighting a US jury verdict that awarded more than 243 million dollars in damages over a fatal Autopilot crash, a case built on internal data that victims' lawyers recovered after the company said the vehicle's data was corrupted, according to an RTS investigation published in April 2026. A federal judge confirmed the award in late February. Tesla can still appeal.

China has taken a different route. On 13 February, the Supreme People's Court stated in a case study that drivers remain responsible for road safety even after activating driving assistance functions, citing a September 2025 Zhejiang ruling in which a driver surnamed Wang was jailed and fined for relying fully on assistance while intoxicated. The court said driving assistance systems "cannot replace the driver as the primary operator." Lower courts must reference the standard.

Put together, the picture is less about whether robotaxis work than about what happens when they meet conditions nobody modelled well. A flooded Atlanta street is not an exotic edge case. It is rain. If a company with 170 million autonomous miles and a decade of testing still has to pause service in four cities to solve it, the industry's expansion targets deserve more scepticism than they usually get.