Elon Musk is seeking to grab eyeballs and headlines with a splashy event Thursday that is designed to impress investors but is short on safety and testing details. Experts say Tesla is far behind Alphabet’s Waymo , which is rapidly growing its commercial robotaxi service.
At Tesla ’s much-hyped “CyberCab” event tonight at Warner Brothers Studios in Burbank, California, Elon Musk is expected to announce a new electric model that will form the basis of Tesla’s futuristic fleet of taxis. Musk is likely to use the invitation-only event, which he has dubbed “ We, Robot ,” to promote the revenue-generating potential of this type of business for Tesla and announce a launch date.
His desire to excite investors and Tesla fans could boost its stock price, which has stagnated this year due to declining sales of electric vehicles. But there is little public information about Tesla’s progress in autonomous driving.
It stopped filing reports with the California DMV, one of the main sources, in 2019 and doesn’t report to outside experts or the media on its progress. There’s also ongoing scrutiny of its Autopilot and Full Self-Driving software by regulators, and known and ongoing technical issues with these suggest achieving Musk’s vision anytime soon would be a stretch.
“Tesla is not a serious contender for real-world applications of autonomous driving,” said Missy Cummings, a George Mason University professor and artificial intelligence expert who has been an adviser to NHTSA on autonomous vehicles and consults with the California Department of Motor Vehicles on the topic. “I think they are at least 10 years away from these types of operations.”
Hollywood stunts aside, the most promising robotaxi program right now is not Tesla’s CyberCab but Waymo’s, which has long been the industry benchmark for autonomous driving. Fifteen years after it began as the Google Self-Driving Car project, nine years since its first fully autonomous test drive ferrying a vision-impaired passenger through Austin in 2015, and with the help of at least $10 billion in investment, Waymo has built a real, if far from profitable, business.
“At the end of the day, it’s a huge opportunity” – Dmitri Dolgo, co-CEO of Waymo
Its robotaxis, with no humans behind the wheel, currently make more than 100,000 paid rides a week in Phoenix, San Francisco and Los Angeles. That number is likely to rise soon, as the service expands to Austin and Atlanta. The vehicles are not faultless, engaging in odd behavior such as inexplicably honking their horns in a Waymo parking lot, freaking out in front of Kamala Harris’s motorcade and occasionally causing minor traffic jams when blocked by double-parked cars and trucks. Waymo has avoided the serious accidents that crippled rival programs from Uber and Cruise, as well as crashes and fatalities linked to Tesla’s partially automated software, though the NHTSA is investigating several collisions and traffic violations involving its robotaxis (those cases resulted in no injuries or fatalities).
Dmitri Dolgov, co-CEO of Alphabet, who has been working to solve the autonomy problem since Google launched the initiative in 2009, is trying to keep it that way.
“Everything we do is approached with security as the overriding principle,” he told Forbes. “That extends to the scaling [of the service] that you see now. It means we’re continually raising the bar on system performance, on the rigor of our assessment, on validation, and on the readiness framework. All of those things are evolving, and that’s how we’ll continue to approach it.”
Bullish estimates for a future robotaxi market are at $50 billion a year by 2030, according to a recent report by Raymond James. Forbes estimates Waymo’s revenues will exceed €2 billion.
Bullish estimates for a future robotaxi market are at $50 billion a year by 2030, according to a recent report by Raymond James. Forbes estimates Waymo’s revenue will top $50 million this year and could approach $100 million as the service expands, but Dolgov won’t confirm that or discuss when it will be profitable. He also won’t discuss Tesla.
“At the end of the day, it’s a huge opportunity and it’s great that this field is attracting brilliant minds and resources,” he told Forbes. “There’s room for different approaches, and at the end of the day if you can solve this problem and accelerate the mission and vision that’s great.”
Human monitors, multiple sensors
Waymo and Tesla have taken very different paths to solving self-driving. Waymo has been running structured testing for years — on its own circuit in California’s Central Valley — and supervised testing in cities, both in the Phoenix metro area and Silicon Valley . It has worked to master specific driving maneuvers that humans take for granted, like making unprotected left turns and navigating around road obstacles, doing them countless times, as well as teaching its AI to recognize road signs, traffic cones, emergency vehicles, and the difference between plastic bags on the street and hard objects. It started with slower, easier-to-maneuver suburban streets with fewer pedestrians and lower speeds before turning its attention to dense urban environments, like San Francisco.
Waymo’s system also relies on remote human monitors who keep an eye on its robotaxis, directing vehicles if road conditions throw them off, interacting with police in the event of an accident or simply answering drivers’ questions. This adds to the cost of its operations, but is key to Dolgov’s goal of ensuring safety, he said.
Waymo also uses much more powerful sensors than Tesla to ensure its vehicles can see and understand as much of the world as possible, including laser lidar for 3D imaging in daylight or darkness, radar, cameras, and audio sensors (important for hearing emergency vehicles).
Tesla, by contrast, collects real-life driving data from its customers through a neural network to train its software to handle any of the road and weather conditions drivers experience. And for passenger vehicles, Tesla only uses cameras as sensors. It’s a much less expensive approach, but one that worries AV researchers.
“Any form of autonomous control is only as good as its sensory input,” says Bart Selman, a professor of computer science and engineering at Cornell University, noting that the Boeing 737 MAX 800 crashes are an example of what can happen if a sensor malfunctions and no redundancy is built in. “I would be surprised if Tesla could deliver fully autonomous, reliable driving with the help of just a camera. It may appear to work well most of the time, but it doesn’t seem robust and safe enough for widespread deployment.”
Tesla’s current automated software, a $15,000 option sold to customers as Autopilot and FSD (Full Self-Driving), has been highly controversial and has never lived up to Musk’s expectations. A detailed review of accidents by the Wall Street Journal, including several that resulted in deaths, found that Tesla’s system has difficulty recognizing objects and obstacles on the road. The lack of additional sensors beyond cameras appears to be key to that problem. While Tesla’s system alerts drivers when to take control of the vehicle, in many cases this may not happen in time to avoid a collision. California, Tesla’s main market, has sued the company for “false and misleading advertising” over its use of the names Autopilot and Full Self-Driving, saying they can mislead consumers about its capabilities.
Broken promises
Musk has a habit of making bold promises and failing to deliver on them, from plans for hyperloops and solar roofs (unveiled at Universal Studios next to Warner Brothers in 2016) to selling 20 million electric vehicles a year by the end of the decade. That extends to robotaxis: In 2019, he promised that Tesla could have 1 million robotaxis up and running by 2020. Since then, he’s clearly been obsessed with turning Tesla into a robotaxi powerhouse, even though 90% of its revenue comes from selling electric cars, the pollution credits they generate, and batteries. And that looks like a high-margin business — if the AI-based driving system is perfected.
“As exciting as the robotaxi concept is, investors hoping it will disrupt Tesla’s finances in the near term will be disappointed” – Karl Brauer, iSeeCars
“Tesla’s value is overwhelmingly autonomy,” Musk told shareholders on Tesla’s earnings call in July. “So I recommend anyone who doesn’t believe Tesla would solve vehicle autonomy to not own Tesla stock. Sell your Tesla stock.” It’s unclear how long it will take Tesla to do what Musk promises.
“As exciting as the robotaxi concept is, investors hoping it will disrupt Tesla’s financials in the near term will be disappointed,” said Karl Brauer, a longtime industry analyst for consultancy iSeeCars. “Getting self-driving cars past regulatory hurdles and onto the road in significant numbers will take time, even if Elon deemed the technology ready for public consumption this week.”
Unfulfilled promises
Musk has a habit of making bold promises and failing to deliver on them, from plans for hyperloops and solar roofs (unveiled at Universal Studios next to Warner Brothers in 2016) to selling 20 million electric vehicles a year by the end of the decade. That extends to robotaxis: In 2019, he promised that Tesla could have a million robotaxis up and running by 2020. Since then, he’s clearly been obsessed with turning Tesla into a robotaxi powerhouse, even though 90% of its revenue comes from selling electric cars, the pollution credits they generate, and batteries. That’s because it looks like a high-margin business — as long as the AI-based driving system is perfected.
Alphabet also doesn’t break out Waymo’s revenue from its Other Bets business, which rose 28% to $365 million in the second quarter, or discuss its expectations for the company, though CEO Sundar Pichai is optimistic.
«I’m really pleased with the progress Waymo is making, a true leader in the space and receiving rave reviews from users,» he said on Alphabet’s earnings call in July.
George Mason’s Cummings is dismissive of Tesla’s program and sees Waymo’s efforts as more comprehensive, but he remains skeptical that any company can completely dominate self-driving any time soon.
A review of test data submitted to regulators by Waymo, Cruise, and Zoox shows that self-driving cars “can drive in ways that are generally technically legal but that sometimes severely violate the ‘rules of the road’ expected of human drivers,” he wrote in a draft of a research paper under peer review that was shared with Forbes. (Tesla is not included in the analysis because it doesn’t share data on “disengagements” — situations in which a test driver overrides the autonomous system.)
Some of the reported accidents “could be the fault of other drivers (such as a truck crashing into the autonomous vehicle on a curve), but humans expect some form of defensiveness and cooperation in the environments where these events occur, and autonomous vehicles do not consistently display these abilities,” he wrote.
Whether it’s Tesla or Waymo, researchers like Henry Liu of the University of Michigan, who runs a self-driving car test center in Ann Arbor, believe the federal government should create a driver’s test specifically for them before they can be allowed on public roads.
Ultimately, Dolgov has his priorities: reducing the costs of his system and not squandering the advantage Waymo has built by expanding too quickly. The first point is being addressed with a so-called sixth generation of sensors, computing system and software that will be used in lower-priced electric vehicles in Waymo’s fleet next year, including Hyundai’s Ioniq 5 utility vehicle.
“We’ve been very intentional with our scaling and deployment,” he said. “We’re intentionally scaling to enable the Waymo driver to learn as quickly as possible and learn from our customers, learn from the experiences and take that forward.”
Still, growth is coming. “Given where we are today, the scale at which we are operating today and the diversity of operating environments, you can find many zip codes where the environment is completely within the operating domain of what we have today,” he said.
