A dedicated model and more iterations of hardware are clear signs Elon Musk isn’t close to making customers’ cars fully autonomous.
(Bloomberg) — If teasing plans for a standalone robotaxi wasn’t a clear enough sign that Tesla won’t be beaming fully self-driving software to its cars anytime soon, Elon Musk just made this disillusionment all the more obvious.
To recap, Musk claimed in April 2019 that Teslas were one year away from no longer needing human supervision and earning their owners $30,000 annually by autonomously driving other people around. Having failed to deliver this 43 months later, the chief executive officer shared a teaser image of a dedicated robotaxi in December.
This was a head-scratcher. Why go through the rigmarole of developing a standalone robotaxi model if the millions of vehicles that Tesla has been building and selling for years have the hardware needed to eventually drive autonomously, as Musk claimed was the case back in 2016?
Of course, that claim didn’t hold up. Tesla has cycled through several iterations of hardware for its Autopilot system since then, even ditching the radar sensors that Musk said years earlier would have prevented a fatal crash.
Tesla caught criticism from car owners a couple years ago for initially trying to charge for retrofitting hardware, then relented. The reason consumers were upset was clear: Musk had already charged many of them for self-driving capability he said the company would deliver sometime in the future, and had claimed their cars already had the hardware that would be needed.
Now, it appears Tesla is going to further test the patience of its customers. During the company’s latest earnings call, Musk spoke at length about yet another new hardware set, referred to as Hardware 4. His message for Tesla owners with Hardware 3 amounted to a drastic departure from past promises.
Here’s what he said:
With respect to upgrading cars that have Hardware 3, I don’t think that will be needed. Hardware 3 will not be as good as Hardware 4, but I’m confident that Hardware 3 will far exceed the safety of the average human. So what we’re aiming for is how do we get ultimately to, let’s say, for argument’s sake, if Hardware 3 can be, say, 200% or 300% safer than a human, Hardware 4 might be 500% or 600%. There will be a Hardware 5 beyond that. But what really matters is, are we improving the average safety on the road? But it is, the cost and difficulty of retrofitting Hardware 3 with Hardware 4, is quite significant. So it would not be, I think, economically feasible to do so.
A few takeaways from these statements:
- There’s now a different end goal for hardware that was supposed to be sufficient for Teslas to become robotaxis, and maybe even rendering the steering wheel obsolete. The objective is now merely for Teslas using automated-driving features with Hardware 3 to be an order of magnitude safer than the average human.
- Musk suggested Tesla will try again to stop offering free retrofits, and maybe even cease retrofits entirely, even as it’s now charging $15,000 for a product marketed as Full Self-Driving, or FSD, that’s been in beta for years and requires driver supervision at all times.
- The new iteration of hardware — debuting on Tesla’s repeatedly delayed Cybertruck — won’t be the last. There apparently will be a fifth generation at some point in the future.
Tesla may well be improving the capability of its driver-assistance systems, as Musk has repeatedly said, but the statistics it’s released purporting to demonstrate the relative safety of vehicles using Autopilot haven’t held up to close scrutiny. It’s also clear that the public will be far less forgiving of crashes involving or caused by robodrivers than those caused by humans, who get into tens of thousands of fatal crashes every year. Just ask Uber, which sold off its self-driving unit years after one of its test cars killed a pedestrian in Arizona in 2018.
While Musk moves the goalposts for Autopilot, Tesla is cashing in more revenue from FSD, even as it continues to tell customers using it to keep their hands on the wheel and eyes on the road. The company recognized $324 million of deferred FSD revenue in the fourth quarter, by far the most ever.
“FSD revenue was recognized purely for rolling out the FSD beta to a wider group of users, whereas in the past FSD revenue was typically recognized when new functionalities were developed,” Toni Sacconaghi, a Bernstein analyst with the equivalent of a sell rating on Tesla’s stock, wrote in a Jan. 26 report. “We worry about the discretionary nature in which Tesla can opt to release deferred revenues (a ‘cookie jar’) in order to improve financial results.”
Days after Bernstein published that note, Tesla lowered its forecast for how much deferred revenue it expects to recognize in the next 12 months. At $639 million, the company’s forward expectation is now the lowest it’s been in almost four years.
In the same regulatory filing containing that disclosure, Tesla confirmed the US Justice Department asked for documents related to its Autopilot and FSD features. Bloomberg reported in September that the Justice Department and the Securities and Exchange Commission were investigating the company’s self-driving claims. Last month, Bloomberg reported that Musk personally directed a video in 2016 that claimed a Tesla car drove itself, and that the SEC was specifically looking into his role in shaping the company’s autonomous-driving claims.
Musk himself conceded a year ago that developing self-driving technology had been much harder than he previously thought. These recent developments suggest Tesla still has a long way ago, giving prosecutors more time to carry out their criminal probe. The company may have previewed what its defense would be in November, when it sought to dismiss a California customer’s complaint that it has misled consumers.
“Mere failure to realize a long-term, aspirational goal is not fraud,” Tesla said.
More stories like this are available on bloomberg.com
©2023 Bloomberg L.P.