According to a credible new report, Elon Musk has reportedly shut down an internal analysis from Tesla executives that showed the company’s Robotaxi plans would lose money and that it should focus on its more affordable ‘Model 2’.
7 Likes
That’s a great piece, but I especially liked this from the comments:
Even if you believe the dubious claims about how possible FSD is, it's just not a great business worthy of Tesla's valuation.
Being competitive and displacing competitors means significantly underpricing existing taxi and ride-share networks. Which limits pricing and margins. This isn't going to be a 70% margin business. It's maybe a ~20% margin business if they get it right.
It also has the capital inefficiency of a manufacturing company combined with worse capital efficiency than a rental car business. It's fundamentally renting a car for time; just in smaller time blocks. Maybe Zipcar is a closer comparison.
However, remember that the primary driver of rental car company income is from the resale of used cars. Cars get used for a few years and are resold. What's the resale market like for a car with no steering wheel? It doesn't exist. Instead of depreciating 30%-50% over time as a fleet vehicle, a robotaxi is depreciating 100%.
5 Likes
In fairness, if the vehicle really is autonomous (a big if, but grant it for the sake of argument), then steering wheel is obsolete and not needed.
3 Likes
another spot-on comment: " **This disaster is what happens to a company when it doesn’t have a functioning board and the investors worship the CEO. It will become a case study in business schools around the world for how to avoid this disaster. "
4 Likes
No, “in fairness”, all vehicles suffer their age at varying rates. When they get older they are traded down for newer. And when they get really older they go into the wholesale market and end up on curbstoners front yards or “Harry’s Used Car Lot.”
They still have value, just not as much. Now take a car whose software is 10 years out of date, maybe with a rusty fender or two, but would still be otherwise drivable by a teenager or someone who can’t afford a “good” used car. But with no steering wheel and software that is likely unsupported, it has literally zero value.
The average 18 year life of a car - in this case will be half that, if you’re lucky. (Who will these supposed fleet owners sell them to, as Enterprise, Hertz, and the rest do after 3 years and the cars are showing their age?)
2 Likes
Tesla is continuously updating your vehicle’s software. My 2020 Model Y has the functionality of a new model – at a $40,000 discount.
intercst
2 Likes
No it doesn’t. It has the old hardware.
2 Likes
We’ll see. A class action lawsuit will likely get me the $1,500 HW4 upgrade for free.
Interesting that Xpeng is upgrading all its existing fleet to FSD at no cost to the owner.
intercst
2 Likes
Of course, FSD is not just for taxis … it is also an attraction to some owners, depending on what it can do.
Why would the software be unsupported?
I’m not following why a 3 year old car with FSD suddenly has zero value or little value.
1 Like
How long does Microsoft or Apple support old software? I have a refrigerator that’s no longer supported. That’s why.
Not 3 but maybe more. Without a steering wheel it’s good for only one thing. And there’s a reason the car rental companies sell into the used market after 3. Maybe tou can get more but you’re sure not gonna get 18 (the average lifespan of a car these days).
1 Like
This is something that I found weird about the supposed internal Tesla analysis referenced above. It wasn’t in the linked article, but other discussions of the Intercept piece said that Tesla boffins calculated that since cars (generally) would be used 5x as much, new car sales would be reduced by 5x. So we would go from 15 million new cars per year to only 3 million new cars.
I didn’t understand that. Under that assumption, these cars are going to get used up much, much faster than the current average lifespan. A car that gets driven 60K miles (or more) per year during the first three years of usage is probably not going to have a useful lifespan of 18 years. The fleet size will go down, but the number of new cars annually produced to keep that fleet in equilibrium shouldn’t change all that much (and might even go up).
Anyway, back to your point - I would fully expect that robotaxis wouldn’t need to be supported for nearly as long as that 18 years. They’ll have a much shorter useful life. I also think that their manufacturers will support the software for quite a while under a subscription arrangement (which is what Tesla modeled their pro forma on, apparently), and probably longer than MSFT or AAPL since that’s going to be a revenue stream for the OEM. What the cars won’t have is that last few $K of value when they move so far down the value chain that they get used as junkers.
2 Likes
Thank you for the well thought answer.
1 Like
Well, not necessarily/ In addition to taxis, it could be used for private transportation if Tesla provided the right kind of support.
1 Like
OK, point taken. It still doesn’t obviate the fact that once the car is too shabby to be used in a rental it has little to no resale value in the used market. (Unless the “new” owner wants to keep paying subscription fees for the right to not have to steer. That surely makes it uncompetitive with other used cars, no?)
I think the idea is that these things will get ridden into the ground in the taxi market. They were projecting that a taxi/TaaS car would end up replacing 5 individual unshared cars, and would translate into about 60K miles per year.
So they’re going to be more like conventional taxis than rental cars. Almost all of their useful life will get burned up in the 3-5 years that they’re in service. By the time they get taken out of the fleet, they’ll have at least 200K total miles - and probably more. There’s probably minimal used car residual value baked into the pro forma, unlike rental cars that will usually end up being sold with only about 50-60K miles.
2 Likes
That’s possible and I hope you get it.
Maybe, maybe not. How about an active person with compromised eyesight?
OK, there’s a small subset of people who will buy a nearly used up beater with no steering wheel (poor people with compromised eyesight) and there are likely a couple of other slivered categories. Overall I would expect the demand for such vehicles to be small, which leads to (almost) no sales which leads to almost no price.
But presumably Tesla will not be getting any slice of taxi rental (since there won’t be any) so will they be providing the software - and support - for free? Or will the dupe have to pay a yearly software license over and above the other vehicle maintenance? I don’t claim to know, but I see pitfalls.
The current indicators are that the initial Tesla cybercab will be geofenced and use remote safety drivers, similar to Waymo.
If that true, then first generation cybercabs will be worth zero on the resale market unless Tesla is able to upgrade the software to the point of true L5 autonomy.
That last part is a heavy lift. And even then, there still might be situations where you want/need to drive the car yourself.
I think Tesla mechanically fastened the pooch using an inclined plane wrapped around a cylinder* It makes more sense for a robotaxi to look like a Zoox. Easy and quick to hop in and out, easily seats four.
Cybercab looks more comfortable if you are settling in for a longer trip. A little more work to get in and out, and what if you are holding bags?
*TMF won’t let me use the common English word for that.
1 Like