I think TSLA is undervalued even if robotaxi is 10years away. As it is today with just GF3, Y and Pickup, Tesla is imo undervalued. Robotaxi is huge, like a $4T market according to ARK:
White Paper: Why Self-Driving Cars Could Change Everything
If there is a 10% chance that Tesla takes 10% of that market, it implies a $40B value. Which is about what I dare to hope for.
However I do think people here really underestimates the probability of Tesla solving FSD soon. Elon, who we should give some credit to having a history of mostly being right eventually seems pretty confident. And his times have been converging on end Q4 for a long while now.......
I chose to quote part of your post, but I like all of it. I agree broadly, and especially with the idea that Tesla is a long ways in front of everybody else, and closer to actual FSD than anybody else.
The place where I diverge with Elon, is I think he's undercalling how hard the social aspects and road situations are going to be to handle.
On the social side, I'm willing to go with the US and Canada as two markets that are reasonably ready to do something like this. However, how ready are people to see cars with nobody in them driving around in traffic? And how ready is society to handle people drivers being particularly bad drivers around empty cars driving around? There's a social side to this that might happen fast - I suspect it's going to be slower.
There's a technical side to this where I'm in complete agreement that the raw technical capability (feature complete is Elon's term, and I think it's a good one) will be ready sooner than later. Sort of like a carpenter has a box of tools that is finally fully populated and ready to get to work building a house.
But a box of tools doesn't make a house overnight. So if we're thinking of FSD as really being REALLY REALLY advanced driver assistance, then I can easily see that in a year or less - maybe even Q1. And we should all rightly rejoice and look forward to counting our gains as investors in the company.
But driver assist is a long ways away from FSD (which to me means cars with nobody in them, going somewhere for their next pickup; and cars with people in them, but nobody in the driver's seat, or particularly thinking about the driving task or expecting to be called on, ever, to be involved in the driving task).
Some samples of problems (common, and uncommon):
- suburban neighborhoods, and other roads, with no lane markings. The lane is implied in my neighborhood because the street is about 2 and a half or 3 lanes wide, but there are plenty of situations where traffic in one direction or the other needs to stop to allow oncoming traffic through. There's an impromptu need for I-go, you-go, and there's no signage or other help on the street. These 'lanes' also wander back and forth from near the sidewalk out to the middle of the street (and a bit beyond) even when 2 driving cars wide.
- suburban neighborhoods that aren't as old as mine, where you've got about 2 car widths of driving total, and with cars parked on 1 side, you've got 1 drivable lane (more I-go, you-go situations).
- Overtaking, or being overtaken, but a Wide Load. I experienced one recently on the freeway where the wide load driver had their right side wheels on or the far side of the rumble strip, and I still had to use the far side of my lane (2 lanes at that point) to get by.
- Construction and other situations where there are humans using signs and hand signals to indicate go / stop, and then figuring out where the lane is when going (it's gonna be .. awkward .. for the car to stop in the middle of these
)
- Out in the country, my mom's house was down about a 1/2 mile of a dirt driveway. In the summer especially, the weeds on each side and in the middle makes it pretty clear where the ruts are and the car is supposed to drive. Sometimes that driveway develops car swallowing pot holes that you need to drive around. And it's nice to avoid the blackberries that are growing out in the lane.
- Again out in the country, I've got a segment I drive reasonably often where multiple times a winter, a tree or really big branch falls down and blocks the road. So far I know it's happened because somebody else got there with a chain saw and cleared it out before I did. The car's gotta recognize stuff in the road that it can't drive through and stop for it, up to and including realizing that the destination can't be reached due to obstruction. (And false positives on this are gonna be really annoying). Meanwhile that same logic has to recognize that the big poof of tissue paper isn't an obstacle and go through it - or at least make the balancing choice between a sudden and unexpected stop and chancing the tissue paper.
- Oh - and don't hit shredded truck tires on the freeway - that's bad.
I don't claim to begin to know all of the situations that stand between sweet driver assist and self-driving. I do claim that we aren't even working on self-driving yet - we're still figuring out the tools that go in the self-driving toolkit and building a sweet driver assist system, that can morph and iterate over time into self-driving.
But you've indicated that self-driving isn't important to your investment thesis (as it isn't to mine), so we can enjoy the awesome driver assist we have today, marvel at how it advances and handles a wider and wider range of tasks, and sit back and enjoy the view while we find out of Elon's right about a year or two, or not.