Tesla’s so-called Robotaxi service in California is being treated by regulators more like a chauffeured car service than a true autonomous cab operation like its competitors Waymo and Zoox.
A top California regulator recently clarified that distinction, saying Tesla does not have the permit required to operate a real robotaxi service in the state. As a result, the company is not required to submit the same safety and driving data to regulators as its rivals.
Pat Tsen, deputy executive director for consumer policy, transportation, and enforcement at the California Public Utilities Commission (CPUC), explained the situation during an interview on The Driverless Digest podcast. She said Tesla’s ride-sharing service operates under a different type of permit, one typically used for traditional car services.
“What they have from us is essentially a charter party carrier permit. It is the same type of permit that a limousine company would get from the CPUC to provide a limousine service. So, in terms of our view of the person who is sitting in the driver’s seat, that is a driver that is not a safety driver,” Tsen said.
Tsen explained that the reason for this distinction is that Tesla’s AV technology is not advanced enough yet.
“So in California, we define autonomous vehicles to be those that are SAE [formerly Society of Automotive Engineers] level three — meaning that the onboard AI system is capable of navigating designated road conditions within an operational design domain on its own,” she said.
Tesla’s system is considered level two, meaning it still requires a human driver to take over at any time.Waymo and Zoox, by contrast, operate at level four, meaning their vehicles can drive themselves in specific, designated areas without human intervention.
I wonder why Waymo and even Zoox were able to get to level 4, while TSLA is stuck at level 2? Like was speculated in the ‘risks of cameras only’ thread, is cameras-only just not capable of getting to level 4?
I asked Google “why hasn’t Tesla gotten a Robotaxi permit in California?”
Google’s response:
Tesla has not obtained a robotaxi permit in California primarily because it has not applied for the necessary driverless permits, relying instead on lower-level, human-supervised testing permits. The company has not logged required test miles with regulators since 2019, while competitors like Waymo have logged millions.
This is bizarre to me. Haven’t even applied for the driverless permits since 2019? I can only conclude that Tesla’s own internal testing data shows that they wouldn’t be able to complete the regulator-required test miles needed to show true autonomy.
Maybe this is why we’ve also seen the disappearance of ‘unsupervised’ robotaxis in Austin. Tesla appears unable to get to L4-autonomy with their cameras-only system.
Who really cares what was discussed if this is the best comment that you can make? This is a total cop-out. I, among many, have questioned for years why Tesla didn’t do formal testing in California (and other states) as others did – getting permit, reporting to the state on testing miles, testing disconnects etc and all the other formalities. Yes all that stuff Elon would claim is a “waste of time”. Yea, right.
But they refused and left their fate in the hands of owners who used their cars like expensive adult toys rather than actually planning and conducting testing the way others did. Right now, we don’t know the “truth” of any real Tesla testing, And don’t you wonder how much farther along they might be if they did approach testing differently?
But it seems right now is that Tesla has been left in the dust of others.
You guys are making the same mistake I made decades ago to the tune of six zeros not counting decimals, concentrating on technology instead of on business. I love technology, most of my investing is in technology, the fastest growing sector of the economy, which can be verified by comparing the “tech heavy” NASDAQ against the market average S&P 500. But I no longer invest in technology qua technology but in what good it does for the companies I invest in. That is the reason for replying “No” to your question.
To give you one example, Tesla beat out Toyota as the bestselling car in the world with the Model Y despite not having L4. L4 does not increase free cashflow. BTW, my last three cars were Toyotas, I loved them!
Now consider what Tesla had to do to beat Toyota (Detroit and Stuttgart couldn’t do it) and it wasn’t getting to L4. In other words, you guys are barking up the wrong investing tree. Invest in technologies that make money for your companies. Right now I’m concentrating on datacenter related companies. Self driving is still years away from making money, it’s too early to invest in self driving so why sweat it at an investing forum?
Because the dream (some would say illusion) of self-driving is what is holding the Tesla P/E at over 300? Pretty sure that is a reason to “sweat it” if you have a chunk of Tesla stock.
If? In Master Plan IV released just a few months ago Elon said that autonomy was central to Tesla’s business plan. Also keep in mind Elon promised robotaxi service to half the population by the end of last year. Next month, Tesla is supposed to start full production of the cybercab, a vehicle without driver controls.
Tesla is going to have to roll out L4 fairly soon or all those cybercabs will be sitting in parking lots. It doesn’t seem plausible that Tesla currently doesn’t want or need L4 capability.
A more logical explanation is that Tesla wants L4, but their current technology is incapable of it and they know it. Hence no permit applications.
This has been mentioned once or twice. There is a second possible explanation. That Tesla could obtain an L4 permit, but doing so would release too much information into the public record about the capabilities of their system. If they “open up the kimono” to the state of California about their system, it could have some serious consequences they’d prefer not to deal with right now.
Part of that would be valuation. Tesla has positioned itself as having an easily scalable, low-cost robotaxi system - if not in hand, then very close to in hand. If the data revealed to CA regulators doesn’t jibe with that, it could be a problem for them.
But I think that a bigger problem would be the hundreds of thousands of folks who bought FSD and their cars when Tesla was saying that they had all “"hardware necessary for full self-driving capability,” which was from about 2016-2024. They’ve already started suing in droves, including a class action lawsuit in CA. As Tesla keeps making hardware upgrades to the robotaxi versions of the Model Y (increasing the capability of the onboard computer systems, adding a front bumper camera, modifying the cleaning capacity of the cameras, and additional specialized telecommunications hardware to allow more reliable teleoperation)….well, those claims become more damaging.
It seems somewhat unlikely that any of those cars will ever be able to drive themselves, so Tesla may have to pay the piper eventually - but timing matters. Attrition will reduce the number of claimants over time. As cars get traded in, sold to third parties, damaged or stolen, etc…you get fewer and fewer folks who made their purchases in reliance on those statements. You’d rather delay handing over the big tranche of performance data to government officials for as long as you can.
I first found out about Elon when he hosted a battery exchange event. I think it was back in 2013. He told a story that made little sense but battery swapping was in style in China. Kandi, which I was invested in, had even developed a robot to do it. I took an instant dislike to Musk who sounded like a circus barker. The next time I heard Musk was at the 2020 Battery Day. By then Tesla was doing quite well and the 4680 battery project sounded very interesting. I started buyingTesla stock on Sept 28, 2020. It has been very rewarding and I have gotten used to Elon speak. It makes no sense to bet against success, repeated success in many fields like helping Ukraine defeat the Russians with Starlink. Geniuses tend to be oddballs.
The Captain
January 19, 2013
Back when i was not impressed by Tesla…
Testa Motors is one contender making a lot of headlines and it has its full contingent fans and detractors. One reason why Tesla is not a disruptive technology is that the price is too high, there is no under-served luxury car market that I know of. Tesla is taking the luxury car market head on, where it might or might not succeed, but even if it does it won’t disrupt the automobile mass market. The small size --so often derided by American commentators-- and the low cost of the Kandi EV are key features in the disruptiveness of the Kandi model.
February 3, 2014
April 18, 2015
EVs have yet to make the grade. Elon Musk has taken some remarkable steps to make it happen like opening up the patent portfolio and trying to standardize batteries for easier adoption, not to mention the superchargers. The future of EVs is still an open question.
Depends on what we are betting on, doesn’t it? Let’s say, Tesla continues to be a successful company and in ten years as a more mature company the P/E has settled down to a still high, but more modest P/E of 25. Which is what Alphabet is now.
For investors to be happy with their returns, the stock price would have to increase by say, 10%/year. Hopefully more than that, but less than that and it wouldn’t be worth it.
So, in order for the stock to be P/25 in 10 years with 10% price growth, earnings would have to grow by 43%/year.
Even with robotaxis and Optimus, that rate of earnings growth seems far fetched, to put it mildly.
It is possible that P/E could remain above 300 for the next ten years. That seems very unlikely as well.
But what happens to the stock price if the robotaxi business never really materializes the way investors are expecting? Every year of delay means those revenues need to be even bigger. And year of delay means more competition. An unlikely scenario becomes even more unlikely.
So yes, I’m going to say they need L4 and they need it fairly quickly. I’m also going to say they don’t have L4 and what they do have isn’t good enough such that they are willing to share the data.
I bet in fancy casinos, not in the stock market. In fact I have developed a Casino Model of selling covered calls to take the money that gamblers, who think of themselves as investors, spend on call options.
Interesting theory; as are other comments in your post that I could have also repeated here.
My own opinion, which is greatly based on independent sources including the community based FSD tracker and that newer Robotaxi tracker is that Tesla remains far from a viable 5 or 6 9s in a March of 9s scenario. But we have no clarity; no doubt greatly due to needing alternative sources of info as Tesla fails in any meaningful transparency. Another issue is that there is a lot of confusion as to what the March of 9s really means. In all cases it is about safety; though it seems there are variations as to exactly what extreme of safety is really meant.
Based on my own reading, some say the March of 9’s starts with the 9 indicating 90%; others claim that the first 9 is that second one prior to decimal which means 99%. Take your pick which one of these you like or that you think could represent current status, but on the FSDTracker, Tesla FSD easily achieves 90% imo but still seems to struggle to get 99% and seems far from 99.9%. That is not very good. I say all this with the caveat that we don’t really know what Tesla achieves in their internal testing.
Couple this area of confusion with the limited knowledge we have of their Austin testing and it seems Tesla remains far from any level of safety and other capabilities which might demonstrate L4 or better. I doubt there would be consequences that would be harmful if they are more forthright with CA or any regulatory body.
I, and I think others too are “past” caring what Tesla is doing as very many believe Tesla has made bad choices.
As a long time shareholder, I remain actually disgusted about many choices they have made.
Tesla actually continues to demonstrate L4 capability in Austin, so that’s not a proper characterization. True, not many vehicles, but you did use the word “demonstrate.”
As for the FSDTracker, remember those miles have a different ODD than Tesla’s taxis, which is different than Tesla’s unsupervised taxis’ ODD. The FSDTracker includes everything from sunshine to heavy rain and snow, roads all over the US, not geo-fenced, at driver-chosen aggressiveness, and include non-professional driver assessments, etc. We can’t map FSD to robotaxi in a March of Nines meaningful way.
As non-employees, we actually don’t know the choice they have and continue to make. I got some inside information recently and have posted it 3 times, but those didn’t stop numerous posts here making incorrect statements and conclusions.
I was talking about a different type of consequences. Tesla was very aggressive with their public statements about the hardware capabilities of the cars they sold between 2016-2024, claiming that every car sold had the hardware necessary for full self-driving capability.
There is a class-action suit filed by buyers who assert those claims were fraudulent, as well as individual suits from other purchasers. There were several hundred thousand sales of cars with FSD during those eight years (and many many more without FSD, but who might be able to prove their purchase decision was motivated by the statements). The potential liability isn’t existential to Tesla, of course - but it would be sizable.
If Tesla “opens up the kimono” about the safety rate of their self-driving program to CA regulators, that will certainly help litigants in their claims. Tesla might prefer to delay that. The population of potential claimants declines over time, as cars get traded in or sold to third parties (who can’t now rely on Tesla’s pre 2024 statements).
I agree with you that they probably aren’t nearly as far along down the “march of nines” as they have led investors to believe, and so they might prefer to keep their actual progress private for that reason as well. But I think the litigation impact might be enough on its own to affect their decision making.
I agree with this; and actually have felt that I have offered caveats that my speculation regarding March of 9s was just that — an attempt to infer safety level based on other sources of info. I fully know it is not a true “apples to apples” compare.
Even though it is less than a rigidly meaningful way to speculate about what Tesla’s actual safety status is right now, I think both public trackers are presenting enough insight into Tesla’s current status for us to infer the problems are far from resolved.
One can see the reasons “why” drivers are having disconnects by drilling down into the FSDTracker. While you are correct about the ODD differences, FSD has been a “drive anywhere, any time” promotion that doesn’t hold up to scrutiny. I doubt Tesla will ever offer a robotaxi that will even allow or attempt that
So what might we see with this 14.3? Haven’t we heard multiple prior claims about how release #Next is going to be? Based on the crowd sourced community tracker none of the effusive talk was very accurate - none look very good yet and many new releases have actually backtracked rather than show improvement.
