Photo: Richard Bord/WireImage/Getty Images
On Thursday, Elon Musk will announce a project he’s been promising for nearly a decade: a self-driving, on-demand electric vehicle he calls the Cybercab. The details so far have been scarce. Despite the similar etymology, it probably won’t look much like the Cybertruck, the geodesic land rocket that Tesla started selling in low numbers last year. What’s important in this case isn’t so much the car’s design as it is the way owners use it — not just as a personal vehicle but as an autonomous taxi that anyone can rent out. He has staked the very future of Tesla on this enormous pivot, and Wall Street has rewarded his vision. But how the Cybercab’s business model is supposed to work — and make money — isn’t clear. As of now, it looks like Musk is taking a well-known, profitable brand and swapping it for an experimental and currently money-burning one, and there’s no real road map for making it a success.
The Cybercab announcement arrives near the end of a tumultuous year for Tesla. The company started 2024 down more than 40 percent from its 2021 peak, then dipped dramatically again until early April. (Musk’s embrace of Trump, enthusiasm for white-nationalist anti-immigrant policy, and recently resolved legal war with Brazil have not been helping). As spring approached, the Cybertruck remained a sideshow, and investors were wondering when Tesla would add a new car to its aging fleet. It was around this time that Musk first revived the idea of a self-driving taxi, which he would later double down on as the very future of the company. “The value of Tesla, overwhelmingly, is autonomy,” he said during a July investor call. “I recommend anyone who doesn’t believe that Tesla will solve digital autonomy should not hold Tesla stock. They should sell their Tesla stock.”
Tesla, of course, is not new to automation. Its vehicles come with the so-called Full Self Driving system (though it is not, in fact, fully automated, and has reportedly been the subject of a criminal investigation). And Musk built what he claimed to be the world’s largest A.I. data center outside Memphis. But his vision to overhaul Tesla isn’t just about a tech company upgrading its software with more A.I. — it’s a complete overhaul of how he plans on making Tesla profitable. The idea of the robotaxi is, broadly, that customers would be able to hail a cab much like an Uber, only for an autonomous Tesla to pull up. Presumably, Musk would get a cut of any fares, taking some pressure off the company as it sells fewer cars.
In broad strokes, this makes sense. Tesla is making less profit per car than it used to now that electric vehicles are commonplace and the Big Three manufacturers are rolling out competitive models of their own. Uber shares, meanwhile, are near an all-time high. So why not just mash the two businesses together? “It’s not about making money today, but making people believe they’re going to make money in ten years,” said Thomas Monteiro, an analyst who follows the company for Investing.com. The problem is in the finer details. Musk is venturing into a new kind of business — essentially chauffeuring — that will use enormously expensive computing power, while relying on pricy technical instruments, and then will likely have to either cover insurance costs or saddle drivers with higher payments.
Driverless taxis are already on the road in Los Angeles, Phoenix, and San Francisco. The dominant company is Waymo, which is owned by Google parent company Alphabet. But despite its rising popularity, Waymo lost $2 billion during the first half of this year, and its operating costs are kept close to the vest. In July, Alphabet committed another $5 billion to the company. And there’s reason to think that these costs would be higher for another company — after all, this is Google we’re talking about. Since robotaxi technology is, essentially, built around the same kind of pattern-matching A.I. ideas behind ChatGPT, you could expect that the massive power and computing costs for operating a fleet of vehicles would run into the billions. Even a company as large as Tesla may not have the scale to bring down costs as low as Google or OpenAI. The Silicon Valley playbook, which has been remarkably successful, is to throw investor cash at expenses, subsidizing a product the public falls in love with. The price of that product eventually falls, but by that point the company that got there first will already dominate the market. (See: Uber.) Musk does not want to cede the streets of the world’s tech capital to Google — especially after he allegedly had an affair with Sergey Brin’s wife.
But autonomous vehicles come with costs that no car company has ever had to pay for. The core hurdle is safety. Musk and other robotaxi boosters will claim that of course Cybercabs are safe, usually more so than humans. But the research isn’t so clear. According to a study published in Nature, automated vehicles were more prone to accidents than human drivers during dawn, dusk, or when turning — in other words, during rush hour and on streets that are more likely to have pedestrians. Who would be on the hook for an accident? According to MarketWatch, Waymo pays the auto insurance for its cars since its technology actually does the driving. Insurers are already trying to figure out how, exactly, they are going to cover robotaxi crashes, which would be far more expensive than traditional car accidents. “Replacing a conventional windshield costs a few hundred dollars; it’s many times more for one infused with sensors. That has implications for carrier operations, and even their solvency,” according to an industry journal.
Tesla’s sales model seems designed to mitigate some of those costs, since California law requires title owners to pay for insurers. But if that’s the case, why would somebody allow their luxury car to be used by strangers as a taxi and then have to assume the cost of anything going wrong? And are there really enough people out there willing to do this that the fares and subscription costs would make up for lower sales? Musk has never really addressed these kinds of concerns, and it’s unlikely that he’ll dive into them on Thursday. But if he keeps wanting Wall Street to throw money at his vision of a driverless future, he will have to explain how the driverless car business is supposed to make money on its own. “Tesla is running out of options, if you look at the bigger picture,” Monteiro said. “Is this going to be the huge game changer that revolutionizes the industry forever? I highly doubt it.”
