Robotaxis do not reduce traffic congestion more than traditional ride-hailing services because they frequently drive without passengers [1].
This finding challenges the long-held assumption that autonomous vehicles would optimize urban mobility. If robotaxis continue to operate with high vacancy rates, the transition to autonomy may fail to alleviate the gridlock facing many U.S. cities.
Data reported to the California Public Utilities Commission by Waymo reveals that robotaxis are empty for almost half of the miles they drive [1]. This practice, known as "deadheading," occurs when a vehicle travels to a passenger pickup location or returns to a staging area without a fare [2].
According to the reports, Waymo vehicles spend close to 50% of their operating time deadheading [2]. Because these vehicles occupy the same road space as human-driven cars, the lack of passengers does not result in a net reduction of vehicle-miles-traveled [1].
Urban planners had hoped that a centralized fleet of autonomous cars could replace private vehicle ownership and reduce the total number of cars on the road. However, the current data suggests that the operational patterns of robotaxis closely mirror those of human-led ride-hailing services, where drivers also spend significant time searching for passengers [1].
The findings highlight a gap between the theoretical efficiency of autonomous networks and the practical reality of current deployments in California [2]. Without a significant shift in how these fleets are managed, or a move toward shared rides, the environmental and traffic benefits of the technology remain elusive [1].
“Robotaxis are empty for almost half of the miles they drive”
The data indicates that the primary barrier to reducing traffic is not the driver's skill, but the logistics of demand and supply. As long as robotaxis operate on a point-to-point model that requires frequent empty repositioning, they will contribute to urban congestion similarly to Uber or Lyft, regardless of whether a human or a computer is behind the wheel.





