Tesla’s robotaxis in Texas are nothing like Waymo’s. They ply within a small geo-fenced area in Austin. Trips to and from airports are currently excluded and Tesla says its service may be limited or unavailable due to bad weather. Most importantly, for the first several weeks at least, Tesla will have a safety monitor sitting in the front passenger seat of the vehicle, while remote drivers also closely monitor each ride as backup. After General Motors’s robotaxi subsidiary Cruise’s operations were shuttered two years ago following a San Francisco incident in which an elderly lady was killed, Tesla is clearly taking no chances. An accident like Cruise’s could severely damage Musk’s reputation and derail Tesla’s ambitious plans to become a global leader in autonomous vehicles, artificial intelligence (AI), humanoid robots and energy storage.
Tesla currently has a market capitalisation of US$1.05 trillion ($1.34 trillion). It produced just 1.77 million EVs last year and is expected to manufacture 2.07 million vehicles this year. It now has a 2.75% share of the total global passenger car market and over 12% share of the global EV market. Tesla sales in Europe and China have hit a speed bump lately and its main Chinese rival, BYD, recently overtook it to become the world’s top EV maker.
The move from EV manufacturer to robotaxi operator will dramatically transform Tesla’s business model as it shifts from selling hardware or EVs to a recurring revenue model with software-like margins. Tesla’s aim is to scale up its robotaxi business by expanding, initially in the US and eventually in Asia, Europe and around the world. There would be over eight million Teslas on the roads around the world by the end of this year. Think of Tesla’s robotaxi business like Airbnb. Anyone who has a home or part of a home to rent can do so on the Airbnb platform. Tesla is betting that with its unsupervised full self-driving or FSD, Tesla owners will happily rent their vehicles out to earn some money rather than have them sit in their driveway or the office car park.
The average American drives his or her car for no longer than an hour every day. That means for 23 hours a day, the car is not utilised. There are two ways the robotaxis could evolve. People will give up their cars and switch entirely to robotaxis because the cost of car ownership, insurance, road taxes, petrol and car parking fees makes them prohibitive. Or they can afford the car but are happy to earn some money renting it out for robotaxi duty.
See also: Meta Platforms weaponises capital to play catch-up in AI race
Vertically integrated player
Musk wants Tesla to be an integrated player that designs and manufactures robotaxis as well as runs a ride-hailing service that competes with Uber and Lyft in North America, Grab in Southeast Asia and Didi Chuxing in China. Little wonder then that robotaxis are critical to the Tesla investment case because once Tesla is able to prove that it can provide a reliable and safe service using a camera-based supervised FSD system, it could dominate robotaxi services on price and scale despite Waymo having the first-mover advantage.
Waymo first introduced its robotaxis in San Francisco in late 2018. It currently has a fleet of over 1,500 cars with plans to add another 2,000 robotaxis over the next 12 months. Tesla is starting with just 16 cars in Austin. Having fewer cars than Waymo isn’t a big issue. Tesla manufactures over 5,600 cars a day, all of which can be converted into robotaxis with FSD software. Yet while Waymo has been on the roads for over six years, Tesla is a Johnny-come-lately in the robotaxi business. Musk has been talking about “switching on” a million Teslas with FSD software for seven years now.
See also: Microsoft says Chinese hackers are exploiting SharePoint flaws
Waymo uses Jaguar’s all-electric I-PACE cars that retail for around US$72,500 in the US. Tesla, for its part, is using its Model Y, whose most basic version costs US$45,000 in the US. Add more bells and whistles and you would be paying much more. On the eve of the Austin launch, Musk reminded followers on his X platform that Tesla is a vertically integrated robotaxi operator. “These are unmodified Tesla cars coming straight from the factory, meaning that every Tesla coming out of our factories is capable of unsupervised self-driving,” he added. Tesla sells FSD software to facilitate self-driving. There are more than eight million Teslas on the road, most of which are capable of working on FSD software. All those cars could effectively be ‘‘turned on’’ and operate as robotaxis.
Here is a brief primer on Robotaxi economics. US$45,000 is the retail price for Tesla Model Y. It probably costs Tesla less than US$35,000 to actually make the car. As a vertically integrated operator, Tesla has advantages that Waymo doesn’t. Moreover, customising a Jaguar I-PACE for robotaxi operations is expensive. Waymos need an expensive array of sensors, radars and Lidar, or light detection and ranging, a method for determining ranges by targeting an object or a surface with a laser and measuring the time for the reflected light to return to the receiver. Tesla robotaxis have no lidars or radars; they use a more machine-learning-oriented approach. Lidar costs US$12,000. The cameras on Tesla cost just US$400. Musk’s idea is that a Tesla car using FSD software should be able to make decisions on its own, more like a human driver might. “Lidar is lame,” Musk has noted repeatedly. “It’s expensive and unnecessary.”
Waymo’s total per-car cost is reportedly just over US$250,000. It competes against a US$35,000 Tesla. As Tesla switches to Cybercabs with no steering wheel or pedals, analysts have gone as far as to predict that prices could eventually drop, probably as low as US$15,000.
Here is how a robotaxi is cheaper than owning a car. Adjusted for inflation, the cost of owning and operating a personal car has not changed since Ford’s Model T rolled off the first assembly line in 1908. A personal car currently costs an average American US$1.10 per mile. A human-driven ride-hailing vehicle costs, on average, US$2 per mile. Tech-focused fund manager ARK Invest estimates that autonomous taxis at scale could cost consumers as little as US$0.25 per mile within 10 years, the price point which will spur widespread adoption.
Data-gathering juggernauts
Think of EVs as data harvesting machines on wheels. They capture and store a huge amount of data, including location details and even the driver’s and passengers’ facial expressions. Waymo topped nine million rides last year, while its main Chinese peer, Baidu’s Apollo Go, completed three million rides last year. Yet, with the largest data lake of all players globally, Tesla’s proprietary data advantage should give it a competitive edge. Tesla is gathering just under 10 million miles of FSD data per day. Compare that with Waymo’s 250,000 miles per day.
Why is gathering data so important for robotaxis? For one thing, it helps train Tesla’s and Waymo’s AI algorithms. In a recent white paper that it published, Waymo mentioned it has noted that with increased data and computation, the performance of autonomous vehicles improves. Though Tesla only introduced its robotaxi service this past week, its cars have been collecting data for years wherever they go.
Sink your teeth into in-depth insights from our contributors, and dive into financial and economic trends
One big issue for driverless cars until recently was regulators, who took years before they would allow state-of-the-art mobility services on the streets. Before his recent row with President Donald Trump, Musk moved swiftly in his role as the head of the Department of Government Efficiency, or DOGE, to convince regulators like the National Highway Traffic Safety Administration, or NHTSA, that they needed to expedite permission to companies like Tesla so they could compete with rivals at home and abroad. NHTSA has cut approval times from a few years to a few months. On June 23, Texas enacted a law requiring operators of autonomous cars like robotaxis to get approval from the state’s Department of Motor Vehicles before operating on public streets without a human driver.
Waymo’s entry last week into Atlanta, in partnership with Uber, will bring additional competition to the state. Lyft Inc, Uber’s main North American rival, is expected to launch its own Atlanta robotaxi service — with safety operators on board — in partnership with Japan’s Toyota Motor Corp-backed May Mobility Inc in August. Another robotaxi start-up, Amazon.com Inc’s Zoox Inc, has also been testing its driverless technology in Atlanta. And then there is Tesla, which is eager to take on the other three with its version of robotaxis next year.
Toyota is not the only legacy automaker that is toying with a robotaxi project. German auto giant Volkswagen, which has been making EVs for the past four years, is getting into robotaxis with Uber as well. The duo will launch their first service in Los Angeles, where Waymo has been operating for the past year. Uber already has a partnership with WeRide Inc to operate robotaxis in the United Arab Emirates and with Wayve.ai in London.
Robotaxis will soon be all over Asia. China has had limited robotaxi service for a few years. Goldman Sachs, in a recent report, predicted there could be up to 500,000 robotaxis operating in over 10 cities in China by 2030. Waymo will start trials in Tokyo this year. China’s Baidu Apollo Go recently announced it is looking to start a robotaxi service in Singapore and Malaysia by the end of this year. Volkswagen is testing its Buzz EV microbus robotaxi in Germany, while WeRide, Pony.ai and Wayve expect to launch their own robotaxis in Europe next year. Uber is teaming up with its ousted co-founder Travis Kalanick to take over the non-China operations of Pony.ai.
Unlike ride-hailing with just one or two dominant players in each market, there will likely be multiple robotaxi operators globally because the aim is not just a taxi alternative but a car replacement. Since Tesla already sells globally and can activate the robotaxi function with its FSD software, Musk’s EV firm will clearly be a formidable robotaxi player globally. “With the largest data lake of all players globally, its proprietary data advantage should give Tesla a competitive edge” against rivals, notes ARK Invest’s Cathie Wood. Yet, bcause mercurial Musk is back running Tesla, it will likely be an eventful ride.
Assif Shameen is a technology and business writer based in North America
