Robotaxis should be a wakeup call for cities

An urbanist take on the driverless future

This morning’s guest post is from Benjamin Schneider, a freelance journalist, Bloomberg CityLab contributor, and the writer of “The Urban Condition” on Substack. He is currently working on a book for Island Press about why American cities have stagnated — and how they can change for the better.  

Matt will be back in your inboxes later today with a few post-vacation thoughts on the election.

---

In San Francisco, robotaxis are like naked guys. The first time you spot one, it feels like a big moment. After that, they fade into the background. 

Until 2022, most of these cars tooled around the city with backup drivers behind the wheel. Then, the drivers began to disappear, and platoons of cars were suddenly navigating the city’s famously chaotic streets all on their own. Yet, it didn’t feel like local or national media were capturing the gravity of the moment. Robotaxis didn’t fit into the narrative that autonomous vehicles were still years, if not decades, away.

Photo by Mario Tama/Getty Images

As a local reporter in San Francisco, I was intrigued by this plainly visible disconnect. In 2022, I wrote half a dozen articles on the rise of robotaxis for the San Francisco Examiner, and more the following year for MIT Technology Review and on my Substack. The more I learned, the more it began to feel like we were sleepwalking into a new era. 

Matt was right when he wrote back in April that robotaxis are underhyped, and that the biggest hurdles facing this industry are economic and regulatory, not technological. 

But his piece neglected some of the most important implications of the rise of robotaxis. This not-so-new form of urban transportation could exacerbate many of the worst side-effects of America’s car dependency. It could also offer a unique opportunity to curb those impacts, and ultimately put all cars in their rightful place on city streets. 

The sneaky rise of the robotaxi

How did robotaxis sneak up on us? It’s not just the stealthiness of Waymo’s purring electric Jaguars. Robotaxis flew under the radar thanks to the hubris of the broader autonomous vehicle industry, and the subtle operational distinctions between robotaxis and other kinds of AVs. The late 2010s was a time of feverish excitement about AVs – US Secretary of Transportation Anthony Foxx claimed that by 2021, “we will see autonomous vehicles in operation across the country.” But 2021 came and went, and instead of driverless cars, the mea culpas started rolling in.

All the while, robotaxi-focused AV companies like Alphabet-backed Waymo, GM-backed Cruise, and Amazon-backed Zoox were quietly clearing regulatory hurdles and racking up driverless miles. These companies have huge advantages over traditional carmakers that aspire to sell autonomous vehicles to the general public. A personally-owned AV must be able to travel anywhere, anytime, in any weather condition. A robotaxi doesn’t have to. These fleets are extensively “trained” in the cities where they’re set to operate. Their service domain can then be restricted to the areas the computer can safely navigate. And fleets can instantly be recalled in the event of bad weather or a software failure. 

So robotaxis glide along. In addition to its robotaxi services for the general public in San Francisco and Phoenix, Waymo has opened waitlists for future riders in Los Angeles and Austin, and is testing vehicles in numerous other cities. Zoox is training its cars in the Bay Area, Las Vegas, Miami, Seattle and Austin in preparation for future commercial service. Elon Musk has announced that Tesla will reveal a new robotaxi service on August 8, four years later than he previously predicted.

Cruise, however, offers a cautionary tale. The company was forced to pull out of California and restructure its business following an accident in which a Cruise car dragged a pedestrian several feet after she had been hit by another vehicle. At this critical stage, every robotaxi company is just one serious accident away from regulatory purgatory. The future of this industry is anything but a sure bet. But betting against robotaxis, or wishing them away, presents risks of its own. If this industry can deliver what it promises, the impacts on cities could be profound. 

The impact of robotaxis on cities

In all of the cities where robotaxi companies currently operate, or plan to, ridehail services from Uber and Lyft are ubiquitous. Public transit is a reasonable option for many trips, particularly to popular destinations like airports and downtowns. Broadly speaking, robotaxis will provide a redundant transportation service. After the novelty factor fades away, the only way these companies can succeed in big cities is if they can compete on price, convenience, and comfort. 

If robotaxi companies can accomplish that, many people would gravitate to this new travel option. Some riders would peel away from public transit systems. Others would choose to make new trips that they otherwise otherwise would not have taken. Many more cars would double park as they pick up or drop off passengers in the busy neighborhoods where ridehail services see the highest demand. 

These three factors, absent any other intervention, are a recipe for significantly increased traffic congestion. 

The emergence of Uber and Lyft in the early 2010s offers a clear precedent. These services worsened traffic in major cities, and corresponded with a significant dip in transit ridership. One study found that 60% of ridehail passengers in big cities would have taken transit, walked, biked or not made the trip had ridehail service not been available. In San Francisco, Uber and Lyft vehicles accounted for an estimated 30% increase in congestion between 2010 and 2016. 

The rise of robotaxis could bring even more severe traffic and labor impacts. This time around, there would be no labor costs — the “other dude in the car” former Uber CEO Travis Kalanick was so eager to remove from the equation — to set a floor on prices or consumer demand. The economic pain and job loss faced by taxi drivers would be multiplied many-fold, as a much larger pool of ridehail drivers would be put out of work.

Over the long term, there would be some benefits to a city crawling with robotaxis. These cars will likely be safer, on average, than human drivers, if they aren’t already. As Matt wrote, robotaxis could accelerate the ongoing parking policy revolution, creating demand for new building typologies with no parking at all, and reducing time wasted circling for a spot.

But that, as far as I can see, is where the urbanist benefits of robotaxis will end. A city designed primarily for transportation by robotaxi is a lot like a city designed primarily for transportation by human-driven cars. Because cars — including autonomous and electric cars — are so inefficient at moving lots of people at the same time, a robotaxi-dominated transportation system would have a relatively low ceiling on the number of people it could transport. Automotive capacity on city streets and highways would remain a paramount concern.

This, in turn, would add fuel to the fire of NIMBYism, making it harder for cities to add the housing they desperately need. Perhaps new neighbors taking up street parking would be less of a bogeyman, but new people and businesses adding to local traffic would remain a fearsome problem. Transit and bike infrastructure would become an even tougher sell, as people take more car trips over longer distances. Already, opponents of transit referenda in Austin and Nashville have pointed to robotaxis as a justification not to invest in new light rail systems.

An urbanist opportunity?

Forward-thinking urban planners like Jeff Speck counsel just the opposite, though. Now is the time to carve out bus lanes and bike lanes and car free streets, avant le robotaxi déluge. “You can allow the growing traffic pressure of a second auto age to ream out your cities anew, or you can decide how much additional space in each street to hand over as demand skyrockets,” Speck writes in the new edition of Walkable City. “The proper answer, by the way, is zero.”

Ironically, robotaxis may offer a golden opportunity to advance the project of a less car-dependent urbanism. America’s gas-tax reliant Highway Trust Fund is a ticking time bomb, as more and more drivers buy electric cars. The clear solution is to adopt a distance-based vehicle miles traveled tax. This need not be an exact replica of the gas tax. It could be a variable congestion charge that goes up or down depending on traffic conditions, Brookings Institution economist Clifford Winston has argued. Such a tax would be far more sophisticated than New York City’s ill-fated zone-based congestion pricing scheme, but could ultimately have the same positive impacts wherever it is implemented. 

Since they lack an existing user base, robotaxis could be subject to these charges with relatively little controversy, before they’re eventually rolled out to all vehicles. Charges would rise when and where traffic is at its worst and transit is at its best. Thus, people would be encouraged to take robotaxis at the times and in the places where they make the most sense: at off hours or in transit inaccessible areas. Charges could also be optimized to encourage first mile/ last mile rides to transit stations. An enlightened policy regime would plow that money back into transit, bike lanes and pedestrian infrastructure — the forms of transportation that can move a lot of people at the same time. 

Another transportation management technology that would help cities adapt to robotaxis is designated pickup and dropoff zones in popular neighborhoods. This would combat the scourge of double parking by requiring robotaxis to pull over in specific areas on each block where the curb is clear.

Cities should have implemented designated pickup and dropoff areas and required Uber and Lyft, as well as delivery companies like Doordash and Postmates, to use them from the start. Already, diminished gas tax revenues are forcing the U.S. to fund roads with hundreds of billions of dollars in deficit spending — figures that will rapidly increase in the coming years without a vehicle miles traveled tax. 

There’s a long history of policymakers reacting too slowly to changing transportation technologies. Robotaxis could help usher in a more proactive transportation policy regime that better serves the needs of cities and their people. Or, robotaxis could create an even more auto-centric transportation paradigm that consigns everyone to spend still more time in the car.