The full $7,500 federal tax credit for Tesla customers ends in less than two weeks, and we’re doing everything we can to try to ensure those who order a vehicle today can take delivery by December 31st and take advantage of the savings.
Per Tesla, that full tax credit was available throughout 2018.
I don’t disagree that for the three years from 2020 through 2022 that no tax credit existed but for the vast majority of all years (2008-2018/2019 and now again in 2023), Tesla was provided a full or partial tax credit. I could have been more clear that I meant during 2019, not all years after as it pertained to partial credit.
Anyone that thinks that the credits did not drive sales is simply being silly. It would be like trying to claim that the recent reduction in sale prices on August 15th won’t drive sales.
That would be (roughly) 40% growth per year from the 1.8 million that Tesla will likely sell this year (more or less).
But how do they get there, even in 2024, given how long it’s taken them to bring new models to market? Can you sell 2.5 million cars if 70% of them are one model? Or 3.5 million in 2025, with the same lack of model diversity?
Almost all of their sales growth has been Model Y - Model S and X are low volume lines, and sales of the 3 peaked a couple of years ago. They don’t have much diversity in their lineup for that level of volume. Can they really sell 1.9 million Model Y next year? That’s an awfully high number for a single car - no one in the history of the automobile has ever sold that many of a single model annually. The Model Y is a massively popular car (it will be the best-selling car in the world this year, in all likelihood) - but does it run into a ceiling at some point?
Had they been able to get the Cybertruck to large-volume production on the timeline they originally projected, they wouldn’t have to rely quite so much on the Model Y. And if it takes them as long to get the “Model 2” into production as the Cybertruck, (four years from unveil to sales), they won’t have any more new models in that 2030 time horizon. Since it’s really hard to imagine consumer tastes being homogeneous enough to make 10 million Model Y’s a possibility, Tesla really has to speed up its timeline for getting new models into production if they have any hope of achieving that goal.
I think Tesla’s biggest problem with ramping up is governmental obstacles.
Oddly enough, Tesla was able to build most quickly in China, of all places. At least initially. Recently, they’ve been told that their requested doubling of the Shanghai facility would not be welcome.
IMO, locating their facility in Germany was a poor move for multiple reasons. With respect to the topic, it’s pretty darn hard to get through all the hurdles there, both from government as well as activists who vehemently oppose every move.
In Mexico, Tesla was promised quick approvals. “Quick” ran out months ago and the 2024 startup is now looking like 2025. And rumors of a factory in India will probably move to an actual announcement soon… but India is famous for foot dragging and if an agreement is made with Modi’s government this year, I wouldn’t expect production until 2026 or 2027.
So yeah, it looks really difficult for the ongoing ramp up. The factory in Texas has a lot of expansion underway, but I doubt it can crank out an additional million units next year.
As for how many of a single model can be produced, we shall see. The next gen vehicle coming from Mexico could well set the new record.
It probably SEEMS like a long time, but it’s fairly typical. The auto industry likes to claim they can bring out a new vehicle quickly… in “X” months, but it’s not true. This may sound absurd, but they’ll work on a new vehicle for years… then after a long time they’ll reach a certain program milestone and claim THAT’S the start. Then they proudly point to how they came out with a new vehicle in 24 months or something. This probably sounds like I’m making it up, but that’s how it works. Why play a game like that? Good question…
Rob
He is no fool who gives what he cannot keep to gain what he cannot lose.
If it has not been said yet, software can be hacked as can hardware eventually if not now. The updates like to Windows 11 are for security and new updated features.
That actually isn’t really what I was talking about. It’s certainly feasible, with enough capital, to get to a point where Tesla has sufficient factories to produce 20 million units in 2030.
The question in my mind is whether they can sell them, based on the likely models they’ll have available during that time frame. Right now, they only have two high-volume production models - the Y and the 3. They will certainly have a third - the Cybertruck - over the next seven years.
But that’s probably it. Given that the Model 2 hasn’t even been revealed, it’s probably not going to be in heavy production before the end of the decade. Which means they’ll have to sell upwards of 10 million Model Y’s in 2030, at least, in order to hit 20 million cars.
I’m just not sure that’s realistic. Whatever the merits of the Model Y (it’s an insanely popular small SUV!), global consumers have different needs and wants for the form factors of their vehicles. Today, across all brands and all models and all drivetrains, global customers “only” buy about 7.5 million small SUV’s. At some point, you reach the limit of how many customers are in the market for a particular form factor of vehicle, regardless of how awesome it is.
Tesla’s projections don’t seem to pencil out unless the Model Y is getting 200% of the market share for SUV’s of that size. Which seems unlikely.
I just don’t think you can sell 20 million units across only three high-volume models. It doesn’t matter how good the models are - there’s just too much variety in consumers’ needs and preferences for that to ever work.
I don’t think they are missing it they are just betting on FSD. If you are the only car company with FSD then how many are you going to sell? I know if they do get FSD that will be the only car I will want.
I can see that. But I remember being told pre-pandemic that the price of a Model 3 was going to triple once released so buy now! Its imminent!
I work for Nvidia. Part of our company is working on FSD. Personally, I’m not holding my breath. And I’m not entirely sold on wanting the feature for me, honestly.
They seem to be slowly falling behind on that point. Both Porsche and Polestar will soon be using MobileEye, a level 4 system. FSD is still a lvl 2 system.
Correction: Porsche will be using Drive, which is 2+ (no idea what the + means) where as Polestar will be using Chauffeur which appears to be lvl 3.
I think there’s a pretty good chance that you won’t see governments allowing people to have autonomous cars unless they have the potential for human operators to take control remotely (like Cruise and Waymo) - which means that Tesla won’t be the only car company with FSD, since those companies already have the tech to do that.
And even if things break the way Tesla wants, and they are the first to reach “FSD,” there’s also a chance they won’t be the only one providing it for very long. Once the general level of computer science/AI/technology progresses to the point where it’s possible for one company to do it - once the problem is solved for the first time - you should see other companies being able to do it as well thereafter. It’s possible in that scenario that Tesla might have a few years as the only game in town, but not longer than that, and possibly much shorter.
Is it really a lvl 2 system? To tell you the truth I do not know exactly where they are at with it. I know they are only allowed to operate at lvl 2 but I think they are much further along than that but are not going to unveil anything until they get to at least level 4. At least that is my assumption.
No - I meant that other companies would be in a position to just develop their own. Once the field has moved to the point where autonomy is possible, it’s not all that likely that only one company will be able to develop autonomy software. Autonomy itself won’t be a trade secret or patentable - so much like almost every other technological advance (“talkies,” color TV, air conditioning, automatic steering, GUI’s, video streaming), if the general level of technological knowledge in the field is advanced enough that one company is able to make the product, others will soon be able to also make that product.
It’s pretty unlikely, IMHO, that Tesla will have the field to itself for very long. I mean, I’m skeptical that they’ll be the first to reach FSD to begin with - but even if they are, other companies will have developed their own within a few years. Even a year or two would be enormously profitable for Tesla (they’ve got a lot of existing drivers they could theoretically sell FSD to) - but it’s not likely to put them in a position where they’re the only game in town for new cars (and thus pull millions of unit sales) for long enough to solve the problem caused by lack of model diversity.
Interesting. If they do plan to license it, such would seem to cut into the idea of them being the only company - especially if one compares it to them signing agreements with just about everyone to use their existing network. That advantage would seem to be rather short lived. While I certainly applaud Tesla for making charging more widely available, it has to have a negative impact on their future sales. I know my wife decided to not buy an EV due to the limitations of the non-Tesla charging network but said she would reconsider once her perferred EV could use the larger network (she wants a SUV larger and higher ride than the Y* but cheaper than the X so she has no choice but to shop other companies). She is going to go with a plug in hybrid in the meantime.
*Y rides about 7 centimeters lower than her preferred hybrid. X runs only one centimeter lower by comparison.
I agree with you Mark.
I bought my M3 in 2018 and got one of the few remaining with the full credit. I think that the Model Y never had a tax credit (until very recently with the IRA) and became the best selling car model in many locations.
(as a software developer, I tend to agree with the idea that once it has been done, lots of others are able to do it)
But…
Look at Google search? Since they showed the best (?) and most profitable way to do it, now everyone is doing it, right? OR how about street view maps, do we have 10 competitors there?
Lots of software needed to make Starlink work, will we have 4 or 5 alternatives?
Google search isn’t enormous because they were (ever?) the only one offering it. At one time, there were at least a dozen major players trying to compete for search - and even today, there are still a few very small alternatives (Bing, DuckDuckGo, Baidu, etc.). Google was able to consolidate market share by successfully competing against alternatives - not by simply being the only option available.
If Tesla’s driving suite is materially better than that of competitors, then of course it might be able to consolidate a lot of market share. There’s reason to be skeptical of that, of course - the global auto market has long been fragmented, resisting consolidation, so even a superior AI driver might not overcome that. Nor do we yet know whether Tesla will even have a better AI driver than competitors - or even if it does, whether consumers will care all that much.
But Tesla is exceedingly unlikely to be the only company technologically capable of offering an autonomous vehicle for very long. And that’s even if it is the first, for certain definitions of autonomy (it’s certainly not the first to get to Level 4).
So I think that’s kind of a longshot solution to Tesla’s model diversity problem, given that they want to get to 20 million annual sales. Toyota’s the largest seller in the world at 10.5 million (2022), but they spread that out across about 30 models over the Toyota and Lexus badges (not counting different trims or hybrid variants or their Daihatsu or Hino subsidiaries), so no single model sells much more than 1-1.3 million units. Tesla’s going to try to move double the volume with a fraction of the models. Are consumers’ tastes homogeneous enough for that to work? We’ll find out in 2025, when Tesla has to sell more than two million of a single model in one year.
It isn’t like other forms of technology. You have to have a large data set to get to FSD. It isn’t like you can put a couple of cars on the road and get there. If you look at Tesla their data has gone Parabolic and that is because of all the cars they have out there feeding their models. The more miles driven the faster they get there.
Is that true? I mean, not that you need a large data set. Just the intimation that you need one as large as Tesla has, or that having one as large as Tesla has gives you an advantage. That’s Tesla’s theory of the case - but is it accurate?
As you point out, Tesla’s data has gone parabolic. But I don’t believe that’s translated into a parabolic improvement of FSD in Tesla’s most recent updates (though obviously since I don’t have an FSD-equipped Tesla that’s not from firsthand knowledge). There’s no guarantee that having more data leads to a materially improved product on the other end - especially if a lot of the data is duplicative or has declining marginal utility. How much extra benefit do you get with data from 10,000 cars on the 405 than 6,000 cars on the 405? FSD is good, but not ready for prime time yet, from what I’ve seen. But perhaps that’s wrong - have users seen improvements commensurate with the parabolic increase in the size of the database?
On the other side, at least two other companies have developed full Level 4 autonomy systems without having as large a data set. And because they’re operating autonomously, they have a much larger set of data about what their programs do in circumstances when they don’t switch off to a human driver.
If they were only to drive on the 405 probably not much, but if they are going to drive all over the world, probably alot. After all, the companies that are now at level 4 are mostly geofenced in and or only at a slow speed. But there is another way this could go to. Maybe when Tesla gives out subscription services for their FSD the other companies decide that they really do not need to have their own, just like the charging network. Then Tesla has a subscription service for most cars on the road. Just another possibility.