The most important fact about robotaxis is also the most obvious: no one is at the wheel. Unlike taxis, per-trip costs can be low because a rider no longer needs to buy a driver's time. Unlike private cars, robotaxis do not sit idle for most of the day, so the investment to build them can be spread across many more journeys. Robotaxis are losing money at present—a trip with Waymo costs roughly a third more than a ride-hailing service on average—but they ought to become much cheaper as fleets are built in bulk and carmakers redesign vehicles that no longer need to accommodate a driver. Tesla is trying to use cameras rather than more expensive lidar sensors.
The average American household allocates 15% of its spending to vehicle ownership. For city- and suburb-dwellers who are not ferocious gearheads, drastically reducing this would be tempting. In the countryside, which may lack the population density to sustain a robotaxi network, things will probably not be too different.
America is home to 1m taxi and bus drivers and over 3m truck drivers—about 3% of the working population. Other potential losers are less obvious: without car accidents there will be less demand for personal-injury lawyers; if people stop buying cars, dealers and used-car salesmen will go. Robotaxis might compete with short-haul air travel and even hotels if some are kitted out with beds.
Cheap urban transportation could cause horrific traffic jams. Congestion is a classic economic externality: the cost of any single car's contribution to a jam is spread across everyone on the road. At present, urban traffic is constrained by the cost of paying someone to drive or the inconvenience of driving oneself. Without such constraints, the result could be brutal gridlock. An economist's answer is to put a price on traffic; presenting a congestion charge as a "robot tax" might make it easier to swallow in America, where road-pricing is formidably unpopular. Self-driving cars are much less likely to break traffic rules and incur fines, meaning something will need to fill a hole in city budgets.
Parking spaces take up a quarter of the downtown of the average American city. Most could be put to better use as housing or offices. On-street parking could become drop-off bays or pavements. Denser and better-connected city centres ought to be an economic boon. Outside city centres, however, robotaxis will probably lead to sprawl, since longer journeys will be more tolerable. Self-driving cars could pull people away from more space-efficient transport such as buses, subways and trains, prompting a "death spiral" where fewer customers mean lower revenues, leading to worse service and fewer customers in turn.
An average working American spends just under an hour commuting each day against eight hours on the job. Turning even a sliver of that into work could boost output appreciably, according to Will Denyer of Gavekal, a research firm. Self-driving cars offer a smoother ride, which should make it easier to get work done on board. Fewer accidents mean not only fewer human tragedies but also lower hospital and rehabilitation bills.
Human drivers tend to bully their robot counterparts; risk-averse algorithms tuned for safety would rather get cut off at an intersection than risk a crash, and drivers know it. Pedestrians can wander across streets without the frisson of fear that comes from jumping in front of a vehicle with a potentially erratic human driver. Driverless cars are also easy to steal from and vandalise.
In San Francisco, a thin majority of residents opposed robotaxis in 2023 when Waymos hit the streets. By late 2025 two-thirds were in favour.
Do not use that foreign word "ideals". We have that excellent native word "lies".