Robotaxis must improve sensor infrastructure to weather climate risk
For robotaxi providers, extreme weather events can interrupt services, damage vehicles, and weaken passenger trust. Implementing better on-board sensors will help, but it will not make robotaxis climate-proof. Providers should pair more resilient sensors with municipal infrastructure. However, sometimes providers may have to accept that the safest service is no service at all.
Flooding creates an operational blind spot for autonomous vehicles (AVs)
Climate change is driving up the frequency of extreme weather events around the world. One particularly high-risk event is flooding. Extreme floods have forced robotaxi providers to shut down operations for significant amounts of time.
For example, in May 2026, Waymo recalled around 3,800 robotaxis after one of its driverless vehicles entered an untraversable flooded road in San Antonio. The company’s safety filing said that the vehicle detected the flooded section of the road but did not respond appropriately. This highlights a serious issue for robotaxis operating in unstable conditions—the ability to detect water does not mean they can determine whether it is safe to drive through.
More: GlobalData’s new AV forecast service
Conditions like standing water, heavy rain, spray, snow, and fog make it difficult for AVs to determine the safety of driving, as they can obscure road markings, potholes, and debris. The robotaxis’ sensors may also struggle with reflections from water, rapid changes to visibility, and the sudden weakening of roads. This means roads can go from passable to completely unsafe quickly and without warning, forcing robotaxi providers to suspend them.
Downtime has a commercial and reputational cost
While necessary for safety reasons, these shutdowns have a complex and serious impact on the robotaxi providers. Firstly, there is the commercial effect: robotaxis that have been pulled from operations do not bring in any revenue.
This is significant in the short term, but the more serious cost may be reputational. If AVs cannot be consistently relied upon, passengers are less likely to book them over human-driven taxis. Consumers may also become wary of passenger safety. This uncertainty matters to regulators as well. Robotaxis have faced a slow road to approval in a small number of cities. This is unlikely to speed up if robotaxis are known for being suspended on safety grounds.
Despite these risks from shutting down in floods, the risks they would face from continuing operations would be far higher. Damage to vehicles could take them permanently out of action, or a person could be seriously injured.
See also: Autonomous Vehicles: Strategic Intelligence
Better LiDAR is not necessarily enough
The technology behind AVs is advancing constantly. Recent advances in LiDAR, sensor fusion, cleaning systems, and perception software will likely improve performance in adverse weather conditions, but they will not remove every risk created by extreme weather. Better sensors improve what the vehicle can see but do not always provide enough information for a safe decision.
LiDAR may help an AV identify an object in heavy rain. However, it may not be able to determine with confidence whether floodwater is shallow enough to cross, whether the road beneath it is damaged, whether a drainage system will overflow, or whether a route will remain passable in several minutes.
As climate change worsens, extreme weather will escalate. This means robotaxi providers cannot just wait for things to improve or rely solely on LiDAR improvements. They must take action.
Collaboration throughout the sector will be key
Robotaxi providers should first invest in sensor resilience. Long-wave thermal imaging may strengthen obstacle detection when cameras are impaired. Ground-penetrating radar can map geology below floodwater to detect changes to surfaces, while acoustic sensing will help the AV system pick up on audio cues. These changes will improve information access for robotaxis. However, relying entirely on the ability of a single vehicle to make these decisions still carries risk in flooding.
To mitigate this, robotaxi providers should consider moving beyond the model of isolated vehicle autonomy, towards incorporating urban infrastructure. This would involve working with local government, road authorities, and emergency services to develop standardized vehicle-to-infrastructure networks. The vehicle could then draw upon hydrological sensors, weather alerts, emergency alerts, and drainage telemetry directly to make robust safety decisions during flooding.
Even then, some conditions will remain too uncertain. Here, robotaxi providers will need to accept the principle from aviation where proactive fleet grounding is standard health and safety practice. The sector will need to agree thresholds for service restriction and fleet suspension that will be adhered to by all. This will require cooperation between operators but could prove the safest option for advancing the technology in a volatile operating environment.
Aoife McGurk
Senior analyst, Strategic Intelligence, GlobalData