Ever since development began on autonomous vehicles, speculation has run rampant about the rise of self-driving taxis, more popularly known as robot taxis or robotaxis. In 2017, outlets from the Wall Street Journal to Forbes were predicting robotaxis and other self-driving vehicles would soon lead to the end of car ownership in big cities. More recently, Fortune Business Insights projected the global robot taxi market would grow from $1.71 billion in 2022 to $118.61 billion in 2031––a 69-fold increase in just nine years.
As I write this in 2025, though, the robotaxi revolution has remained little more than a ripple. Uber, Waymo (Google), Cruise (GM), Zoox (Amazon), and Tesla are among the companies working on US robotaxis, but roll-out has been slow, to say the least. Waymo has the biggest current fleet, but they are only available in Los Angeles, San Francisco, and Phoenix. Development in China is a little farther on, with companies like Apollo Go and WeRide starting to make waves—but their footprint is still minuscule compared to the population size.
Despite the potential advantages of a robot taxi fleet, they appear to still be years, if not decades, away from widespread use. What are the reasons for this super-slow adoption of robotaxis? The answer has several layers and intertwined parts. Today, I’m discussing four reasons why robotaxis have been slow to catch on.
Technology Issues
Although self-driving technology has come a long way since its introduction, it still has a long way to go. WardsAuto (part of Wards Intelligence) points out that, among other things, there is work to be done in making robotaxis better at real-time sensing and data processing. Furthermore, they can’t just operate well in pristine conditions. Autonomous taxis need to account for rain, snow, and other poor weather. They also need to handle constantly changing traffic patterns—not to mention the sometimes unpredictable behavior of human drivers.
Suffice it to say there have been growing pains on this front. In May 2024, the National Highway Traffic Safety Administration (NHTSA) opened investigations into Waymo due to a series of incidents and Zoox because of multiple rear-end collisions after its vehicles stopped unexpectedly. The previous October, a Cruise robotaxi in San Francisco hit a woman and dragged her 20 feet before stopping, which led to the California DMV revoking Cruise’s operating permit. (It didn’t help that Cruise refused to turn video footage over to regulators.) These are just a few of the many public black eyes that technology hiccups have given robotaxi developers. Even if net incidents are less than human-driven vehicles, it creates a perception that’s difficult to shake.
Development and Operating Costs
Money is also a big drawback. In the long run, industry experts believe that robotaxis will cost much less per trip than traditional ride-hailing and ride-sharing. In the short run, though, there are significant financial hurdles to developing and rolling out self-driving taxis on a large scale. This was seen when Cruise unveiled its first driverless car, the Origin, in early 2020 amid estimates that it could cost up to $400,000 because of specialized equipment and creating production lines. In 2024, the New York Times reported Waymo’s fifth-gen robotaxis were still reaching a $100,000 manufacturing cost plus technician, storing, charging, and insurance expenses—costs that are passed on to the consumer.
Fortunately, prices appear to be coming down, and Tesla is adamant that its Cybercab, which is planned for a 2026 release, will be under $30,000. In the meantime, though, the high costs are a hindrance to a wide roll-out. One evaluation of Waymo services in San Francisco between late 2023 and mid-2024 found an average rate of $11.84 per mile plus a 2% city surcharge. By comparison, estimates put the average Uber ride in the city at $3.25 per mile in 2023 dollars. That gap must be closed significantly for robotaxi use to increase.
Lacking Infrastructure
The self-driving tech isn’t the only limiting factor in robotaxi adoption. These vehicles require fast cellular connections to receive maps, traffic information, and other data so the internal systems and sensors can operate correctly. If local network infrastructure isn’t up to snuff, it can cause robotaxis to malfunction or not work at all. Even large cities aren’t necessarily prepared for an influx of robotaxis, as demonstrated in August 2023. One night after Cruise received approval from California to expand its San Francisco robotaxi fleet, 10 of its driverless cars were rendered inoperable because they lost connectivity, causing a traffic jam in North Beach.
Cruise publicly blamed the connectivity and operating issues on higher cellular demand from a local music festival. This underscores how much bandwidth is necessary to power a growing robotaxi squadron on top of existing daily and peak needs. This is especially true in densely populated cities like San Francisco (only New York City has a higher population per square mile).
The struggles are magnified when you consider the San Francisco Bay Area houses Silicon Valley. If the biggest and most tech-savvy regions in the US can’t consistently handle a relatively small number of robotaxis, what chance do smaller cities with inferior infrastructure have? The rise of 5G will help, but if anything, bandwidth will need to progress at an even faster rate than the vehicles.
Trust Among Passengers
Finally, even if the technology is in place, it won’t matter if potential passengers aren’t accepting it. As Billy Joel once sang, it’s a matter of trust—and plenty of people still don’t trust driverless technology. Even driver assistance technologies and partially-autonomous features built into everyday vehicles have experienced backlash. I’ve written previously for AutoInfluence about concerns of driver assistance features getting too intrusive; although I ultimately concluded the answer was no, it’s impossible to ignore that many people think they are.
This extends to driverless cars, with a Pew Research Center study released in 2022 finding 63% of US drivers were reluctant to ride in an autonomous vehicle. Along with already-discussed reasons like costs and safety, respondents also expressed worry that self-driven vehicles could harm the environment since they make traveling by car more accessible, thus increasing usage. Notably, people under the age of 50 and holding college degrees were more receptive, though opinion was still a split decision at best.
On a related note, concerns have been raised about the loss of jobs and the potential legal liability issues from robotaxis. As of November 2023, Uber alone had 6.5 million people driving for them in the US at least part-time, and Lyft added another 1.3 million. That’s a lot of people who need to find other income sources as robotaxis phase their jobs out. The Journal of Institutional Economics has also pointed out there is currently no generally accepted liability policy for robot accidents—so if two robotaxis crash into each other, who or what is responsible? This is another potential barrier to building trust and acceptance.
The Future of Robotaxis
Nevertheless, companies are forging ahead with plans for robot taxis, even if not at the rate once envisioned. Clearly, though, there are impediments to overcome for robot taxis to go mainstream. And through it all, despite all the hypothesized benefits, the Rand Institute points out there is still no official framework to appraise robotaxi safety, making it difficult to evaluate. Cruise and Waymo have released data claiming reduced collisions versus human ride-hailing, but one must acknowledge they have a vested interest in such a result.
I’ll leave readers with this: in a 2023 Chinese robotaxi study published in Systems, researchers concluded that “performance expectations and effort expectations positively influence usage intentions” and vice versa. In other words, if potential customers perceive that robotaxis offer a better experience, they’re more likely to use them. So, as technology, costs, and infrastructure continue to improve, robotaxi expansion may still come down to altering the public view as much as anything—so hopefully, these companies have good marketing teams.
