That is incorrect.
I stand by this statement. It’s pretty simple Albaby, either I will be right or wrong, since you are taking the other side of that, you will be right or wrong. Now I just have to sit back and watch. I like my chances. Of course I expect you to try to put all kind of qualifiers around that statement to hedge your bet. 
Andy
How is that incorrect?
If Tesla were the last automaker to get to Level 5 autonomy (to take an extreme hypothetical), I think you would agree that limping across that finish line wouldn’t necessarily increase Tesla’s market value very much (certainly not by $1.3 trillion). Nor if Tesla decides that its AI is good enough to be Level 5, but no regulators agree and it’s not legal to actually use the product anywhere. Nor if Tesla finally gets to Level 5, but many years down the road after Level 4 products are ubiquitous and being used for 98% of trips. After all, if all the other automakers are already selling a Waymo-powered Level 4 car that can drive itself anywhere in any major city, Tesla’s not going to be able to make as much money selling Level 5 cars.
Do you not think that Tesla can build the Mexico factory by 2025 and then release the (so-called) Model 2 a couple of years later? I have to say I have no idea what the plans really are but this certainly seems possible. Even if it takes 3 years after the new plant is open they could be ramped up by 2030-ish time frame.
Compare to how long it took to build the Shanghai factory (just over a year?) and the Berlin (18 months plus 6 months of bureaucracy ) and Texas factories (about 2 years?)
Mike
Your incorrect because it isn’t logical. Nobody is going to want a level 4 car if they can get a level 5 car. Just like nobody wants an Iphone 2 when they can get an iphone 14. It doesn’t matter if Tesla is last, because after all, you already argued that whoever is first will only have that lead for a short period of time. So if that is correct last isn’t going to matter. With Tesla selling the most cars now at level 5 that will be exponential.
Andy
I really do question how many people even want Level 4. Seriously. I know very few people saying “I really hope my next car can drive itself”.
Take a 3000 mile trip with a bunch of back seat drivers and then we will talk.
Andy
No - I was working off the point that Rob had made upthread, that the Cybertruck probably won’t be a very high volume vehicle, which I agree with. Maybe 500K per year. So the three high volume models would be 3, Y, and “2”.
It will probably be even more dire. The “Model 2” might end up being Tesla’s third high-volume vehicle - but when? If they’re as slow as Cybertruck, and don’t start production until 2028, then they’re barely selling a million of that model in 2030 no matter how good it is. If they’re faster, maybe the Model 2 ramp is up to 2-3 million by 2030 - but that still leaves Tesla getting more than 80% of their sales from just the 3 and the Y, if they’re in the ballpark of 20 million. And if they don’t unveil the “2” early next year, even that’s a tough prospect.
Is that at all feasible? No matter how awesome a car is, can it really support sales of that magnitude?
Albaby
BTW, to quote Philip J. Fry, “There’s no such thing as 2.” My own guess is that the cheap “Model 2” is contingent on Tesla having completely solved autonomy. In their presentation, they showed only a fraction of the lowered cost coming from better manufacturing - the rest was from redesign of the vehicle. I suspect that they’re assuming that they can finally take out not just the gear stalks they’ve been value-engineering out of the car, but all the controls altogether - no more steering wheel/yoke or gas pedal or brake pedal or lane turn buttons or anything else. I suspect they’re years away from FSD, so I actually don’t think they’re unveiling the “2” any time soon.
People aren’t going to pay much more for a Level 5 car if they can get a Level 4 car more cheaply that does everything they want it to do. Most people don’t actually buy the very most advanced or very highest quality version of whatever they’re buying - because those versions are typically much more expensive. That’s why Tesla sells so many more of the 3 and Y than they do of the S and X. If Level 4 is “good enough,” it will be hard for Tesla to sell Level 5 both at a premium and in volume.
That doesn’t mean they won’t make money. Just not necessarily “an additional $1.3 trillion in market cap” money. If there are already lots of cars on the market that come with a full self-driving option, adding one more to the mix that has just a small amount of additional self-driving coverage isn’t going to be able to charge much of a premium off the going rate for that option.
And it certainly isn’t going to be enough of a differentiator to be what allows Tesla to sell 8 million of a single model. If the Model 3 and Y were the only self-driving cars on the market, I can see the possibility that Tesla could move 16 million of them in 2030. But if most cars on the market at that time already have nearly-complete Level 4 autonomy option available, Tesla’s FSD isn’t going to move the needle that much.
Iphone 2 vs 14 is quite disingenuous. Compare instead 13 (which unlike the 14, was available all last year) to Samsung.
The 13 (at 5% of sales) was the best selling phone last year, followed by 13 Pro Max, the 14 Pro Max, and then the Galaxy A13.
Seems like people still displayed a discriminating preference even when there was a demonstrably “best” available.
Note even the most popular phone was still just 5% of worldwide sales. 1.5% of all sales were for the IPhone 12.
You really are underestimating what Level 5 is going to do to Tesla and how far it will drive their Revenues. That is great for the people invested in Tesla, but not so great for the people that believe you.
Andy
Actually what you are stating is disingenuous. Comparing Level 4 to an Iphone 13 and Level 5 to an iphone 14 is completely wrong. The difference between Level 4 and 5 is going to be dramatic.
Andy
I did not compare 13 and 14, I compared 13 to Samsung. I specifically stated that 14 was not out the full year. I went onto to compare 13 to 12. Heck, you can compare the 14 to the 12.
By December, the 14 was the #1 selling phone yet the 12 was still being purchased. People were still deciding to buy something that was cheaper even when something much better was available. Lvl 5 vs 4 will not be 15 years better (the iphone 2 came out in 2008).
Perhaps you could explain what you think that will mean.
A system that is fully Level 4 solely because it is geofenced is basically indistinguishable from a Level 5 system within that geofence. You can’t self-drive outside the geofence, of course. But if the geofence is big enough to encompass everywhere (or virtually everywhere) you might want to drive, then there’s no material difference whether your car is Level 4 or Level 5.
So if a competitor gets a Level 4 system that covers most of the country’s urban areas and the highways between them, most consumers won’t see much practical difference between what that Level 4 system can do for them and a competing Level 5 system. Sure, you might not be able to self-drive 100% of the way to a remote vacation cabin - but you can self-drive for 99% of your trips (and most of the way to that remote cabin, until you leave the highway).
In a market with a bunch of solid Level 4 competitors, Tesla would still be offering a superior product. Level 5 is still better than Level 4. But the size of that performance gap wouldn’t be especially big in that scenario. Maybe big enough to generate some additional pricing advantage for Tesla? Perhaps - though if every car is self-driving in nearly every situation already, will people care all that much? Satellite cell phones provide more complete coverage than regular cell phones - but if you’re traveling to places outside of normal cell coverage that often, you’re going to be fine with the ordinary geographic range.
Yes it will. One is geofenced with a steering wheel and gas pedal. The other can drive anywhere and will no longer have a steering wheel or gas pedal. So yes it will be 15 years better because if it can drive in any condition anywhere you want to go it will be dramatic. So dramatic that people might decide to not even have a car anymore and just call one up to take it where ever they want to go.
Andy
Interesting how this argument is in direct opposition to the one related to range anxiety, i.e., that while most driving might be local, a lot of people make an occasional trip.
Might also note that having Miami on the list of supported places does no good if one lives in Nashville. Might take quite a while to cover all urban areas. And, I suspect Europe there is a lot more driving to other nearby areas.
Even once lvl 5 tech is available, I think we are a VERY long way away from removing steering wheels and go pedals. If you think range anxiety is keeping people away from EV, wait until people have to cope with the possibility of being stranded, even in the rare case of lvl 5 failure (which will most certainly happen at some point).
Personally, I have no desire to own a level 5 vehicle that does not provide me the option to take over the driving in case of failure. I doubt I am alone in that feeling. And, if I have a steering wheel and a go pedal, then lvl 4 geofenced is not that big of a step down.
YMMV.
So you think that having islands of level 4 coverage is going to beat having a car that can travel anywhere? Good luck with that. If people get a level 5 car for the same price as a level 4 car do you think they will really settle for level 4? Because that is what Tesla is going to do to them. Then everyone of the companies will come to Tesla to license their FSD because they will want it also. So yes the licensing will drive their market cap up 1.2 trillion dollars and they is conservative.
Andy
Not only that when a hurricane comes and you are geofenced in you might wish you could get to somewhere else. LOL
Andy
Not really. The car can still go everywhere - it just can’t self-drive everywhere. That’s why PHEV’s don’t present range anxiety concerns. Even though they can’t get everywhere on their electric battery, they can still get everywhere an ICE can.
Certainly. I suspect we’re many years away from a comprehensive Level 4 system - and that we’re many years away from a Level 5 system.
As for Europe, I’m not so sure. One of the reasons why I always thought EV’s would get adopted more quickly in Europe than the US was because Europe doesn’t have nearly as much long-distance inter-city automobile travel. Many countries have both good systems of intercity rail and good intracity transit. Combined with higher petrol prices, European drivers are more likely to use their cars for mostly local travel, and hit the rails and buses if they’re going to be farther from home.
Apple sells 7 models of phone, from a 12 to a 14, with variations for “Pro.” There are many sites doing business selling everything back to a 6, so apparently not everybody buys the most expensive anything, always.
I’m with albaby on part of this: it isn’t the size of the dataset because that won’t all be downloaded into each car. It’s the training, and the ability to make the AI “think”. Watching ChatGPT, they’ve taught it how to “word”, but not yet how to think or evaluate. I don’t need a huge dataset to drive a car, I just need to know how to follow the lanes, to anticipate errant drivers, to not drive into a pedestrian, and so on. I don’t have to have seen every pedestrian ever to make the leap from “a” to “all” pedestrians.
For training the big set is important, but once you get to a certain size the marginal improvement from doubling the set is pretty minimal. Double it again and you get almost nothing for it. And at some point software is replicable; look how many “calculators” are available for your phone. Or word processors. Or spreadsheets, for that matter. The number narrowed because of the dominance of Microsoft and Apple, not because it was hard to make the software behave. Heck, how many AI’s are already out there in the world?
Is driving in Miami so different than Nashville? You’re thinking the AI has to have a map of Nashville. I’m thinking it needs to know how to drive. GPS will take care of the streets, the AI will take care of the driving.