Tesla's recent announcement that it is bringing underwriting in-house for its car insurance unit generated enthusiasm about the prospect of taking data-driven auto insurance to the next level. But I have a different take.
I agree that what Tesla is doing is the next logical step in having sensors in cars feed data about driver behavior directly to underwriters to improve their understanding of the risks. In time, that path could take us to a world where behavior is being monitored in real time and feedback provided to drivers in the form of higher or lower premiums. Encouraged to be careful, those drivers could make the roads safer for all of us.
But I don't think Tesla is the right, well, vehicle for that next step, for a host of reasons.
I'll explain.
Tesla CEO Elon Musk has frequently overpromised and underdelivered over the past decade, primarily on the prospect for fully autonomous driving but also on insurance. After he launched Tesla Insurance in 2019, he said it could soon account for 30% to 40% of Tesla's value. Based on Tesla's market cap today, that would imply a value of $330 billion to $440 billion. For comparison's sake, Allstate has a market value of about $50 billion. Anyone think Tesla Insurance, with a $42 million net underwriting loss for the first nine months of 2024, is worth six to nine times Allstate?
Musk has said Tesla owners should get discounts on insurance because Tesla Autopilot is so much safer than human drivers, but analysts say he fudges the numbers on accidents, and a recent study found that Teslas actually had more accidents per vehicle than any other make in the U.S. last year, as they were in 2023. Starting with bad assumptions doesn't exactly lend itself to good underwriting.
If your publicly stated starting position is that you should offer discounts, and your drivers are no safer or perhaps even less safe than those insured by others, you're in for a rough ride. Perhaps Musk has hired great people and perhaps they'll be very careful as they bring the underwriting in-house, but caution isn't part of Musk's personality or the company's. And we've seen how hard it can be to get the formula right even at seemingly prudent startups such as Root and Metromile.
Musk has been known to power his way through problems before, including at Tesla, but he's CEO of five companies at the moment... and has a day job in Washington, DC that appears to be all-consuming. So he's unlikely to wade into Tesla Insurance any time soon.
Besides, Tesla has bigger problems these days than its insurance operation. Sales fell for the first time last year, even though the market for electric vehicles is growing rapidly and even though Tesla had posted sales gains of 38% and 40% in the two prior years.
Tesla's stock has soared more than 40% since the U.S. presidential election on investor hopes that Musk's $278 million of contributions to the Trump campaign and vocal support via Musk's social media platform, X (formerly Twitter), will translate into success for the business. But the tenor of the Trump administration is, in fact, very much opposed to electric vehicles. Last week, Senate Republicans introduced bills that would not only repeal the $7,500 tax credit for electric vehicles that was Tesla's lifeblood in years past but would charge a $1,000 fee on every new EV sold.
Musk may have other political problems, too. Suspicion is growing that the sales slump partly stems from Musk's hard-right turn politically, which has angered lots of people in what was considered to be Tesla's sweet spot: well-off consumers worried about climate change. An article in Wired says the potential trouble extends to Europe, where Musk has waded into politics in support of right-wing parties in England and Germany. Wired says Tesla sales dropped 13% in the European Union in 2024, and the decline may be accelerating.
"Last month, Norway—where EVs overtook internal combustion vehicles in total market share in 2024—recorded a biting 38 percent slump. At the same time, Tesla sales in France fell by a thumping 64 percent. And it gets even worse: In Spain, Tesla sales plummeted by 75 percent.... Tesla registered only 1,277 new cars in Germany in January, a year-on-year drop of 60 percent.," Wired says.
I also think he's going to have a major problem come June, by which time he's promised to roll out a fleet of robotaxis. Based on the lofty stock price, I may be in the minority. But I've been following autonomous vehicles closely since starting on a book on driverless cars that Chunka Mui and I published a dozen years ago, and I can't believe the technology is ready.
Waymo has been cautiously rolling out its robotaxis for years, based on technology that includes radar, lidar and cameras. It's only operating in geofenced areas that it's mapped in excruciating detail. Yet Musk says he'll be able to put out a fleet that uses only cameras and that can go anywhere.
I just don't see it.
Now, he's plenty good at distracting people with promises that he'll colonize Mars or whatever if we just give him some time. And he does have X/Twitter to broadcast his messages, as well as a supporter in the Oval Office. But I still believe he'll have to do major damage control. If history is any guide, he'll make some small gesture and promise that the robotaxi fleet is imminent -- he's been making such promises about Tesla Full Self-Driving since 2015.
I still believe in the basic vision behind what Tesla is trying to do with its insurance arm, including the move to bring underwriting in-house. Actual data on driving behavior is obviously far more helpful than something like a credit score when evaluating risk, and other recent attempts to supply insurers with more data face pushback. GM's OnStar has been banned from selling driver data to insurers for five years, and Arity has been sued for allegedly collecting data via cellphone apps, without warning drivers that they were being monitored.
Insurers such as Progressive will obviously still have access to driver data that they obtain directly from customers. But the hoped-for breakthrough in integration between car makers and insurers will have to wait, and I suspect Tesla won't be the first to get there.
Cheers,
Paul
P.S. Here's how not to sell extended warranties on cars.