When I mention self-driving cars to someone outside the tech bubble that is the San Francisco Bay Area, they’ll often ask, “Do you really think they’ll become a reality?” To which I respond, “They already are.”
Waymo, for instance, operates a 24/7 robotaxi operation in parts of the Bay Area, as well as a handful of other cities. The autonomous vehicles are practically everywhere you look here, provoking slack-jawed tourists to pull out their phones and take videos. Personally, hailing a robotaxi to run errands or meet up with friends has become so routine that I no longer bat an eye.
At the same time, the limited reach of Waymo and other self-driving companies evokes a quote from speculative fiction writer William Gibson: “The future is already here – it’s just not evenly distributed.”
Self-driving cars, long a sci-fi staple, are becoming a tangible reality for more people, albeit gradually so. Despite recent expansions for companies like Alphabet-owned Waymo and Amazon-owned Zoox, the road to developing autonomous vehicles is a long and winding one — so much so that even major players like Apple and General Motors have pivoted from their endeavors in that space, with GM saying earlier this month it’ll no longer fund its Cruise robotaxi venture.
Those departures illustrate the challenges of building and scaling autonomous driving technology: it’s expensive, competitive and, not surprisingly, there are many regulatory hurdles.
Still, self-driving companies are forging ahead in 2025. Waymo plans to expand to Atlanta and Austin, Texas, through a partnership with Uber and is launching its first international testing in Tokyo. Zoox aims to open up to public riders, starting in Las Vegas. And startup Avride, which has also partnered with Uber to deploy its autonomous vehicles and delivery robots, hopes to launch a robotaxi service in Dallas in 2025. Lyft, too, is teaming up with autonomous vehicle companies like May Mobility, aiming to make its fleet of self-driving Toyota Siennas available to riders in Atlanta starting next year.
Watch this: This Robotaxi Looks and Drives Like No Car You’ve Ever Seen Before
As robotaxi companies gradually expand their coverage areas, they’re likely to remain on the slow and steady route. That means we probably won’t see the proliferation of self-driving vehicles in a large number of cities in the upcoming year but rather more measured growth, such as Waymo expanding its fleet or Zoox slowly ramping up its presence, says Anirudh Bhoopalam, an analyst at Lux Research.
“I wouldn’t expect any big revolutionary developments in 2025,” he says. “To achieve any kind of profitability, you need scale. And if you’re trying to scale so quickly, the probability that something bad happens is higher; that’s what happened to Cruise.”
Barriers like tighter tech regulations in some parts of the world and the general public’s hesitance toward self-driving cars can also slow down progress, Bhoopalam notes.
“There’s a large part of the public that’s excited about autonomous vehicles, but at the same time, there’s a large part that’s very scared about it as well,” he says. “And it really depends on which bubble you’re in.”
Safety over everything
Waymo has been expanding its reach in recent years, while touting its commitment to safety.
Aside from regulatory hurdles, there’s also the considerable challenge of convincing the public that self-driving cars are safe. That’s why practically every piece of communication from robotaxi companies touts the safety of their respective vehicles and technology — especially in comparison to human drivers.
A Waymo data hub published in September states that after driving over 22 million miles, its self-driving tech was involved in “73% fewer injury-causing crashes and 48% fewer police-reported crashes compared to human drivers.” A 2022 Cruise report states that “there is no ambiguity that human driving mistakes are one of the most substantial factors causing roadway injuries and deaths.” And in an open letter last year, Zoox’s chief safety innovation officer noted that, “Data from the National Highway Traffic Safety Administration shows that 94% of crashes are caused by human choice or error.”
This messaging isn’t surprising, given the string of incidents in recent years involving the nascent technology.
Cruise was suspended indefinitely in California last year after one of its driverless vehicles struck a pedestrian, who was first hit by a human-driven vehicle. Waymo’s vehicles have also been involved in a handful of high-profile collisions, including one with a biker in San Francisco and another with a towed pickup truck in Phoenix. (The company recalled and updated its software to address the issue.) And in May, the National Highway Traffic Safety Administration launched investigations into both Waymo and Zoox for incidents in which company vehicles behaved erratically. That includes Waymo vehicles colliding with parked vehicles and sudden braking on Zoox’s test fleet, according to the NHTSA.
Still, these companies have all suggested that driverless tech could actually curb deaths and injuries on the road by ensuring vehicles are doing exactly what they need to, without distraction.
For those of us who have taken a robotaxi, fears about safety may seem inflated. Once you get over the initial weirdness of a vehicle driving itself, the experience becomes exceptionally ordinary, and it just feels like a regular ride. But people “tend to overestimate the maturity of the technology,” says Phil Koopman, associate professor of electrical and computer engineering at Carnegie Mellon University and autonomous vehicle safety expert.
“People are very quick to accept things that seem well-behaved,” Koopman says. “You could have one good ride, you could have 100 good rides, you could have 1,000 good rides and you still have no information about safety, because safety is up in the tens of millions of rides.”
Even one accident can feel like one too many, especially with all the lofty promises about safety from autonomous vehicle companies.
On the other hand, people who haven’t hitched a ride in a robotaxi are likely to be spooked by negative headlines and second-guess the technology altogether. According to a AAA survey from March, 66% of US drivers have expressed fear and 25% feel uncertain about fully self-driving cars. It’s understandable, as I had my reservations the first time I rode in a Waymo, too. (Seeing a steering wheel turn by itself is not normal.) After a few minutes, though, I felt myself relax as the vehicle seemed to know what it was doing, even at four-way stops and while navigating among bikers and pedestrians.
But, Koopman notes, “People overreact to bad headlines; that’s also what people do.”
Companies will likely spend the months ahead continuing to focus on advancing their self-driving technologies, especially across a range of climates like rainier areas (Zoox is conducting testing in Seattle, while Waymo will soon expand to Miami). They’ll also continue to rack up driverless miles to tout the safety of their fleets.
Developments in artificial intelligence could help expedite that process, especially for companies in earlier stages, Bhoopalam notes. AI can help “generate millions and millions of scenarios” in a simulated environment, helping to reduce the sheer volume of testing on public roads (which will of course remain critical).
“That should, in theory, speed up the time between development and deployment,” Bhoopalam says. “But scaling is a different beast altogether. Waymo is scaling now, but it took them many years to get there.”
Expanding access through partnerships
Nuro is one company developing driverless tech for personal vehicles, ride hailing and commercial use. It’s already partnered with companies like FedEx, Kroger and Uber Eats to transport goods.
The introduction of robotaxis raises questions about the fate of ridehailing apps like Uber and Lyft – which themselves shook up the taxi industry when they made their debut. To help mitigate any negative impact, Uber and Lyft (which have each taken a stab at developing their own self-driving platforms) are partnering with companies like Waymo and Avride at an early stage. Uber is also teaming up with Wayve to develop advanced driver assistance and self-driving tech, with the goal of eventually adding Wayve-powered autonomous vehicles to Uber’s network.
Offering autonomous rides directly within the Uber and Lyft apps can not only boost revenues for the ridehailing companies and help maintain their relevance if and when robotaxis take off, but it can also make summoning a driverless ride more enticing for people who don’t want to clutter their phone.
“As small a barrier as installing a new app may seem, people are just more inclined to use what is available in their existing app,” Bhoopalam says.
Another big name hoping to make a splash in the self-driving space is Tesla, which unveiled its Robotaxi prototype at a flashy event in October. Details on the technology itself were sparse, but CEO Elon Musk shared an ambitious goal of kicking off vehicle production “before 2027.” Even more ambitious: his goal to begin rolling out fully autonomous, unsupervised driving in Texas and California with the Tesla Model 3 and Model Y next year.
Experts are skeptical.
“Building hardware is not predictive of success,” Koopman says. “The software is the only thing that matters.”
Zoox co-founder and CTO Jesse Levinson also shared some unfiltered thoughts about Tesla’s ambitions during a TechCrunch Disrupt panel in late October.
“If they had technology that worked, I’m quite confident they could go through the regulatory steps. The more fundamental issue is they don’t have technology that works,” Levinson said, separating Tesla’s so-called Full Self-Driving from truly autonomous systems. (Musk, not surprisingly, was not pleased by this assessment.)
But it’s not just the industry giants that want a piece of this lucrative pie. Startup Avride has been doing testing in Austin with a safety engineer in the vehicle for several months, and plans to launch a robotaxi service in Dallas sometime in 2025, through a partnership with Uber. First, though, it’s teaming up with Uber Eats in Austin to deploy its delivery robots, which will later expand into Dallas and Jersey City, New Jersey. (Avride also operates its delivery robots and conducts autonomous vehicle testing around Seoul, South Korea, and is also expanding its delivery bots to Japan.)
Nuro is another company tapping into autonomous deliveries and ridehailing — along with developing driverless tech for personally owned autonomous vehicles. On the delivery front, it’s partnered with companies like FedEx, Kroger and Uber Eats to transport goods. The company also plans to license its Nuro Driver technology to third parties who want to develop autonomous personal vehicles and rideshare operations. Nuro currently operates its fleet in Mountain View and Palo Alto, California, as well as in Houston, Texas, though members of the general public can’t yet hail a ride.
“There are far fewer potential AV (autonomous vehicle) players in the space than there were a couple of years ago,” Dave Ferguson, Nuro’s co-founder and president, told me during a ride along in Mountain View in mid-December. “What that means is that some of the other players in the ecosystem are more motivated to partner with the remaining AV players. … There’s more motivation to make sure that everyone that is interested in playing their role in the ecosystem has the partners they need to be able to execute on that.”
The Bottom Line
Zoox has been testing its robotaxis in Foster City, California, as well as Las Vegas and San Francisco. The pod-like vehicles don’t have a steering wheel, driver’s seat or pedals.
Ultimately, what the autonomous vehicle race comes down to is money.
“The companies that have the resources to be funded for 5 to 10 more years will make it through,” Koopman says.
There will undoubtedly be road bumps along the way. Robotaxis have received lukewarm reception in some areas, with multiple reports of vandalism to vehicles from Waymo and, previously, Cruise.
In places like San Francisco, there are efforts like the Autonomous Vehicles and the City Initiative, a UCSF-led collaboration between businesses, policy leaders and academics to mitigate concerns around issues like traffic and clean energy. Current and past sponsors include companies banking on the rise of autonomous driving like Waymo, Uber and Lyft.
Perhaps, in the end, the benefits will outweigh the drawbacks. People may gradually become enamored with the idea of having a ridehailing vehicle all to themselves or hopping in their very own autonomous vehicle and not having to pay attention to the road. It’s possible members of the general public will find that they actually do feel more secure with a machine behind the (metaphorical) wheel.
“Once people see that these vehicles can perform well and make their lives easier,” Bhoopalam says, “they will be more accepting of it.”
For self-driving companies, that appears to be a challenge they’re eager to meet, full speed ahead.