Why did Waymo stop robotaxi service in Atlanta and Texas this week?
Table of Contents
Why do Waymo cars keep entering floods? It’s a data delay. Discover why robotaxis trust weather alerts more than their own sensors and what the recall means.
Key Takeaways
What: Waymo suspended robotaxi service in five cities and all highway routes.
Why: Navigation software prioritized delayed weather alerts over real-time sensor data, leading vehicles into floodwaters.
How: Nearly 3,800 vehicles are under recall while engineers update software to fix perception gaps and information latency.
The Alert Latency Paradox: Why NWS Data Cannot Save Robotaxis
A robotaxi doesn’t look at the clouds to see if it’s about to rain. Instead, it waits for a text message from the government. This reliance on the National Weather Service (NWS) created a dangerous “blind spot” during a recent storm in Atlanta. While the sky opened up and flooded the streets in minutes, Waymo’s software was waiting for an official flash flood alert that hadn’t been issued yet. By the time the data caught up to the reality on the ground, an unoccupied Jaguar I-PACE was already stranded in an intersection, trapped for an hour until the water receded enough for a human to take the wheel.
Standard industry logic suggests that autonomous vehicles are safer because they use a suite of advanced sensors to “see” the world more clearly than humans ever could. However, the Atlanta incident reveals a counter-intuitive reality: Waymo’s system architecture actually trusts external administrative data over its own real-time physical perception. Even when onboard cameras and LiDAR are looking directly at a submerged road, the car may continue driving if the official digital “map” or weather bulletin hasn’t flagged the area as high-risk. This hierarchy of information—prioritizing a remote data feed over the “eyes” on the bumper—is exactly why these vehicles keep mistaking dangerous floodwaters for navigable puddles.
The Infrastructure of a Recall: 3,791 Vehicles Under Scrutiny
This isn’t a one-off glitch. In San Antonio, a similar navigation failure resulted in an unoccupied robotaxi being swept into a creek. These repeated failures forced Waymo to issue a voluntary recall through the National Highway Traffic Safety Administration (NHTSA) covering its entire fleet of 3,791 vehicles.
The company attempted a software patch to limit vehicle movement in areas with an “elevated risk” of flooding, but the Atlanta stranding happened after that update was supposedly in place. The current technical remedy is mostly a series of geographic restrictions rather than a breakthrough in how the AI perceives liquid surfaces. Because the software still struggles to tell the difference between a shallow splash and a car-swallowing flood, Waymo has taken the drastic step of suspending service entirely in five major hubs: Atlanta, Austin, Dallas, Houston, and San Antonio.
Beyond Weather: The Compounding Safety Record
While flooding is the current headline, it is only one part of a growing list of “edge cases” the technology cannot yet solve. Federal regulators at the NHTSA and NTSB are currently investigating multiple safety concerns, including a low-speed collision with a child in Santa Monica and more than 25 instances where robotaxis illegally passed stopped school buses.
Beyond safety risks, the vehicles are also creating bizarre urban disturbances. In one North Atlanta neighborhood, residents watched as a line of up to 50 driverless cars invaded a quiet cul-de-sac. Eight of the vehicles became physically stuck while trying to figure out how to turn around, leading one frustrated neighbor to place a “Step2Kid” sign in the street just to deter the confused AI. When the software hits a logic wall it can’t climb—whether it’s a flooded street or a dead-end road—the cars often just stop, creating traffic jams and requiring human intervention to clear the path.
Technical Outlook: The Path to a “Final Remedy”
Waymo is currently working on what it calls a “final remedy” for its all-weather navigation flaws, but for now, the fleet is pulling back. The company has temporarily banned its robotaxis from freeways in Los Angeles, San Francisco, Miami, and Phoenix while it tries to integrate “technical learnings” into its software.
The goal of a driverless future is to remove the human from the loop, yet these incidents show that humans remain the essential safety net. Whether it is a journalist being rescued by an Uber after her Waymo refused to move, or a technician having to physically drive a recovered vehicle out of a flood zone, the “autonomous” system still leans heavily on manual help when the weather turns sour. Until these machines can trust their own sensors as much as they trust a weather bulletin, the expansion into rain-prone climates remains on ice.