The Near Future of Self-Driving Cars
Why Waymo won for now, and what’s coming next
By now, you’ve probably read about the theatrical “We, Robot” event that Tesla hosted late last Thursday evening on a movie studio lot near Los Angeles, where the company unveiled its latest inventions: a self-driving sedan prototype named “Cybercab,” along with a shuttle van concept and an updated version of Tesla’s humanoid robot, Optimus.
Tesla CEO Elon Musk reportedly gave a presentation that was high on grand promises — the Cybercab “could” cost less than $30,000 and “probably” will go into production in 2026, Musk said — but low on technical details, business plans, and regulatory concerns. There was zero elaboration on how Tesla plans to make the major leap from advanced driver-assistance features to fully autonomous vehicles, or whether it plans to run a robotaxi service or outsource it to a ride-sharing partner. Even the supposedly AI-powered futuristic humanoid robots were quickly debunked as being teleoperated by humans and thus in no way autonomous or intelligent.
Needless to say, the market did not react kindly to such empty theatrics. Per Bloomberg:
The underwhelming event sent Tesla’s shares tumbling as much as 10% Friday in New York, the biggest intraday decline in more than two months. They were down 7.6% at 12:29 p.m., wiping out $58 billion in market value. The stock had soared almost 70% since mid-April, largely in anticipation of the event.
At a time when Gen Z already think Tesla is no longer “cool” and Tesla’s EV market share falling below 50% in the U.S. as new entries make headway, one could easily understand why the company would opt to put on such a glitzy show to try to impress Wall Street and gain some momentum. But it proves to be a misguided move that further damages investor confidence in the company.
Another factor behind Tesla’s pressure to perform, of course, is that Waymo is beating Tesla to the punch when it comes to bringing self-driving cars to the market. While Tesla has long offered a driver-assistance system called Autopilot, which can steer, brake and accelerate its cars on highways, it has spent at least four years trying to develop its “Full Self-Driving” (FSD) system, to no discernable results.
In contrast, Alphabet-owned Waymo has already made robotaxis a reality in several U.S. cities, with a fleet of about 700 cars operating in San Francisco, Los Angeles, Phoenix and Austin, Texas. Here is what Alan Ohnsman wrote for Forbes:
After 15 years of R&D, more than $8 billion of investment and multiple pilot programs, Waymo’s robotaxis have become a business, booking more than 50,000 rides a week in the three cities. Assuming an average fare of $20 per ride, annual revenue should top $50 million this year. It was less than $1 million in 2022, according to a Pitchbook estimate — representing growth of about 1000%. Even with modest growth in the four cities in which it plans to operate by the end of 2024, Waymo could be looking at annual revenue of hundreds of millions of dollars within another year or two.
Why Waymo Won, For Now
While Waymo’s Robotaxi fleet will not be taking over all U.S. cities any time soon, it is undeniable that 2024 has turned out to be an inflection point for self-driving technology. And so far, Waymo is winning. And the company’s early success stands in contrast to the lagging performance of Cruise and Tesla.’
2024 has turned out to be an inflection point for self-driving technology. And so far, Waymo is winning.
In hindsight, Waymo’s success over Cruise can be largely attributed to its more measured approach to scaling its operations. Although Waymo has certainly not been immune to safety investigations and minor incidents, taking a cautious, methodical path has allowed it to avoid the high-profile safety incidents that plagued Cruise. According to a safety analysis report that Waymo published of the over seven million miles driven by its driverless vehicles, Waymo’s driverless cars saw 85% fewer injury crashes and 57% fewer police-reported crashes per mile than human driver baselines.
So, while Cruise expanded aggressively into multiple cities, leading to accidents like the San Francisco case where a robotaxi struck and dragged a pedestrian, Waymo prioritized gradual growth and safety, focusing on refining its technology before pushing into new markets. This careful strategy has helped Waymo maintain a better safety record and build trust with regulators and the public, giving it a competitive edge in rollout.
In contrast to this strategic distinction, Waymo winning over Tesla at the moment entails a more nuanced, technical reason. Waymo’s self-driving cars rely on a combination of cameras, radar, and lidar (aka laser-based sensors) to create a 3D view of the environment, making it capable of detecting objects and navigating complex urban environments in various lighting conditions. Tesla, on the other hand, has primarily relied on cameras and software for its FSD system, which has been less effective in achieving full autonomy, further hindering the company’s ability to keep up with its robotaxi competitors.
What Comes Next
Of course, we are still in the early stages of this self-driving revolution, and there’s still plenty of room for potentially Tesla to catch up. As analyst Ben Thompson noted in his latest Stratechery column:
Musk has been over-promising and under-delivering in terms of self-driving for existing Tesla owners for years now, so the jury is very much out on whether current cars get full unsupervised autonomy. But that doesn’t change the fact that those cars do have cameras, and those cameras are capturing data and doing fine-tuning right now, at a scale that Waymo has no way of matching.
The Tesla bet, though, is that Waymo’s approach ultimately doesn’t scale and isn’t generalizable to true Level 5, while starting with the dream — true autonomy — leads Tesla down a better path of relying on nothing but AI, fueled by data and fine-tuning that you can only do if you already have millions of cars on the road.
In short, he argues that Tesla’s camera-only approach may be more conducive to scaling self-driving cars and robotaxi services in the long run because it focuses on affordability and data-driven finetuning. Waymo’s method, while highly effective for safety and precision, faces greater challenges when it comes to cost and scalability.
One potential caveat in this argument is that, given time, the cost of lidar sensors could also gradually come down as Waymo and other lidar-enabled robotaxis start to scale and prompt the market to produce more lidar sensors. If that came to be, then affordability and scalability would not be such a hindrance for Waymo. While Elon Musk has long been critical of lidar, the fact that Tesla bought over $2 million worth of lidar sensors from Luminar in May seems to indicate a potential shift in its approach as well.
Waymo, of course, is driving full speed ahead. To strike the iron when it’s hot, the Alphabet company has extended its existing partnership with Uber to bring its robotaxi to more markets, starting with Austin, TX, and then Atlanta in early 2025. Soon, Uber riders in these cities will be able to book Waymo’s driverless cars exclusively through the Uber app, unlike in San Francisco and Los Angeles, where Waymo’s own app is used for bookings. Uber’s established presence in ride-hailing helps Waymo grow its business without having to develop its own customer acquisition strategy from scratch.
Clearing regulatory concerns remains a major roadblock for robotaxi services. Different states, countries, and even cities have varying regulations when it comes to autonomous vehicle testing and operations. This patchwork of regulatory environments creates challenges for robotaxi companies to launch and scale their services across multiple regions.
In addition, regulators are also rightfully sensitive to the myriad of social and ethical issues surrounding autonomous vehicles, such as job displacement for human drivers, equitable access to robotaxi services, and the opaque decision-making processes within AI systems in dangerous situations. Balancing these societal concerns with the push for technological progress can slow regulatory approval, further delaying the rollout of autonomous vehicles.
Compared to the rather cautious and deliberate approach of the U.S. regulators, robotaxi is fast-becoming a reality in China. As Reuters reported in August, at least 19 Chinese cities are running robotaxi and robobus tests. Among these cities, seven have approved tests without human-driver monitors by at least five industry leaders: Apollo Go (backed by Baidu), Pony.ai (backed by Toyota Motor), AutoX (backed by Alibaba), WeRide (partnering with Uber), and SAIC Motor (the largest of the “Big Four” state-owned car manufacturers of China). In contrast, Waymo is the only U.S. firm operating unmanned robotaxis that collect fares.
The competitive market naturally begets cut-throat prices; In Wuhan, for example, it’s possible to travel six miles in a driverless taxi for about 50 cents, as CNN reports. As with most innovations, cheap pricing lowers the barrier for consumer adoption, which helps scale the operation and further brings down the cost. From here on, how China handles the ripple effects of rapidly deploying robotaxi services will likely set precedents for regulators around the world to consider.
International expansion of Chinese robotaxi providers is also a key factor in how the global market will play out. Although Baidu is planning to expand Apollo Go to global markets outside China through local partnerships, as Nikkei Asia recently reported, it seems unlikely that Chinese robotaxi players will be allowed to operate in the US in light of the US Commerce Department’s announcement in September that it plans to ban Chinese self-driving car software. Interestingly, Tesla released an AI roadmap in early September that laid out the company’s plans for its FSD system, to be available in China and Europe in the first quarter of 2025, pending regulatory approval. So perhaps it will at least have a chance to get to those non-U,S. markets before Waymo.
