TSLA: Scenerios and Ranges

[candide]

Hey, let's make a spin-off thread about TLSA in terms of future cash flows, rather than leaving this in ERE2, future of AI (which I think there is consensus that self-driving cars will happen).

Where we last left off:

13% in TSLA and headed higher.

I have stated somewhere in these parts that TSLA will need competitors in the FSD/Robotaxi industry to stave off Federal anti-monopoly laws. Much the same way Microsoft propped up Apple in the PC market when Apple had existential concerns, TSLA will require a few bit most likely regional players. Maybe let San Francisco putter around in Waymo's until everyone gets dizzy constantly take right hand turns. I said it's a winner take most market and TSLA will no doubt be the AI hostess with the mostest.

Hey look at us, getting some common ground. My turn to say "I already said that": viewtopic.php?p=303760#p303760

Where I thought about Uber as an example of defying commoditization in the U.S. market, and so I put TLSA at having a top range of 60% of the mature robotaxi market.

What's your top range on this "enough to avoid anti-trust" 70, 80, 90%?

Um, as to your "right hand turns," you do get that 1. right hand turns are more energy efficient and 2. Waymo is still a proof of concept that competition is hot for geofenced robotaxis good enough to be "geezer-mobiles."... Can't grandma get most of her errands done (or you know, go to a duck pond) in the same metroplex?

I would love to be able to write as well as read with my eyes while some other entity gets me around the OKC metro. It just so happens that we had solved this problem three generations ago and kept it solved for one generation, before we leaned into car culture.

Also, the robotaxi direction is just in complete tension with selling high-margin luxury. I know Uber is some prior for why there is lock-in on a brand, but they got there by losing money on rides. That's not how Elon rolls. He likes green fields, and frankly getting government subsidies while doing so. (research for the last point using GPT, plus bonus discussion).

[candide]

... And yes, the possibility of Tesla losing all government subsidies should be factored into a DCF analysis of the stock. Since I don't own any outside of in an index fund, I leave pegging those probabilities and run those numbers as an exercise to the reader.

It would also seem that the set of worlds where US regulators force Tesla adoption has narrowed further.

[I normally don't read news, but have it came up as I did some research for my stock positions, so if there was some dramatic irony here the last few hours, well, that's just fun].

[chenda]

It's like an episode of Skins. They might kiss and make up in due course.

[candide]

Let's just say one of the operators is known to be purely transactional in the best of times and otherwise pretty vindictive, and the other is on a lot of ketamine. A lot of stuff was said today that it is hard to see the relationship coming back from.

So I'm going to keep with ranges here...
I think 10% chance of "make up." I think there is at least as good of a chance of investigations being opened/reopened, whether fair or not.

Strongest chance is more contracts being lost than currently on table.

[7Wannabe5]

It's like an episode of Skins. They might kiss and make up in due course.

On the other hand, not all ENTJ and ESTP pairings work out well. Many of these couples ultimately break off their relationships due to their many differences and an inability to compromise.

ENTJ and ESTP are strong personalities who rarely back down from disagreements.
Both types are impatient, which may affect how they interact.
ESTP’s love for risk and adventure conflicts with ENTJ’s need for structure and planning.
These two strong personality types don’t back down easily, especially during conflict. Both strongly feel that their opinion or way of doing things is correct while everyone around them is wrong. This stubbornness can lead to many disagreements that turn heated and remain unresolved. ESTP’s carefree attitude and risk-taking behavior can lead to constant stress for the more disciplined and cautious ENTJ. Their mutual arrogance and impatience towards each other are often why these two may decide to go their separate ways.

[Seppia]

The whole Musk enterprise may be facing intersecting times ahead

Gift article from Bloomberg, don’t know how many clicks allowed
https://www.bloomberg.com/news/articles ... lagQDfa0Ko

[candide]

I don’t plan on following Tesla very closely as I rejoin the workforce, so I am going to keep this big picture.

1. Core business -- luxury EVs.

Multi-month sales slump. Halo effect on brand gone. EU and China almost certainly lost forever... Responding that the cars are nice and that they are the best selling EVs is irrelevant, as even if we froze current sales (rather than saw them shrink further), they would not justify current stock valuation.

Anyone want to do odds of U.S. EV credit going away?

One possible attempt to revive luxury EV is the metaphor of AirBnB. Go ahead and pay more because we’ll offer a button that will turn your car into a small business... Well, are the luxury cars going to be cheap taxis or expensive taxis? Either way, I don’t think affluent people actually like sharing cars. But the cheap price-point taxis will almost certainly pick up wear and tear and smells. There is a reason why Lexus is separate brand for Toyota, and it is not used for taxis.

But even if the luxury cars are kept as a separate tranche of taxi, then that limits the market greatly, and I think still tarnishes luxury. The point of wealth is not to share.

2. Robotaxi!

This is going to be a commoditized, lower-margin space, with competitors already in play. Tesla’s refusal to use LIDAR can be seen as the stubbornness on Musk’s part, or perhaps a cost-cutting measure. It almost certainly trades off with safety, which could hurt adoption, if not getting through regulators.

And then there is fleet size. If the idea is to replace 60% of all drives in the U.S. economy, then the fleet required would be massive. And that will require huge capital expenditures -- again in a low-margin/competitive space. This is the other end of the contradiction of “AirBnB but for cars” ... if the fleet is only people currently owning Teslas, then it won’t be big enough, and other competitors will get ahead in adoption. And if Tesla is going to own its own fleet to try to capture vertical profits, then it is going to be a CAPEX nightmare.

3. Not robotaxi, roboeconomy... to the moon!

This is the only moonshot on the table, as you’re soon going to have to buy xAI stock to get your Grok exposure on. I was going to make a steelman case for the Optimus stuff, but here’s what Musk has to say on the subject:

https://www.businessinsider.com/what-el ... 025-5?utm_

I think we feel confident in getting to one million units per year in less than five years, maybe four years. So by 2030, I feel confident in predicting one million Optimus units per year — it might be 2029," Musk added.

A million robots... 4 years from now, probably 5? And this is from the guy who over hypes everything. There are billions of people working on stuff.

Why would I be more bullish on acceleration than the accelerationist’s accelrationist?

[Sclass]

I’m having trouble understanding how $4.20 cab rides can add up to significant profit. I estimate the car will take a year to pay for itself given energy cost, cleaning and tires. After that you’re making a hundred dollars a day after insurance, cab medallion, vehicle registration. Waymo charges 3x that and it’s hard to see how they’ll be profitable at that level. I guess the way this is analyzed is take an uber fare, subtract out driver expense and then you have robotaxi. Perhaps the business model involves a low interest loan to purchase the car like a rental property. It’s interesting to see the Waymo CEO squirm in her chair when she is asked to discuss their profitability roadmap.

It’s a great tech and I’m deeply invested. But I think the technology will have to go a lot further than robotaxi to justify the engineering costs.

It’s not hopeless. There is demand from the trucking industry, food delivery, mobile vending machines (no driving a route to refill items), postal vehicles, earth movers, buses, licensed self drive for AI impaired automakers. I’m just saying the robotaxi business model doesn’t make much sense if you try to justify its value based on fares.

A front lidar from China (used on Didi AI car) costs only $100. A bunch are currently under production in response to high dollar lidar sets from Velodyne which was the gold standard for self drive. It uses a cheap dvd laser, plastic optics and a cheap galvonometer. Texas Instruments sells cheap chipsets to do the time of flight measurements. Complete junk but it will tell you if there is a solid object twenty feet in front of the car. Saying you don’t want at least one lidar based on cost doesn’t make sense in 2025. I believe the Tesla AI people believe the lidar is a crutch and will ultimately result in a dumb system. Kind of like how Figure AI refuses to use teleoperation on their humanoid bot. You can but it will be used as a crutch. I may be wrong about the motives but I’m quite sure cheap 120 degree lidar is available for $100. Cost is not an excuse.

What’s also really interesting is in the robotaxi videos I’ve seen at least three Teslas drive by the robotaxi that have a big roof rack on top that looks like it is carrying a bunch of instruments on the roof. Maybe starlink hubs, maybe lidar/radar mapping to support the cars. Who knows. But it isn’t my imagination, there are model y cars driving around the same geofenced areas with something that looks like a roof mounted lidar on top.

[zbigi]

What always makes me curious is that no one solved the self-driving train problem yet (apart from trivial cases). I get that the commercial payoff is not nearly as big as with cars, but, given how supposedly close we are to FSD and how difficult cars are compared to trains, Tesla or Waymo should be able to bang out a self-driving train in a proverbial weekend. It may be the case that all the capable players are going after the big reward and ignore easier goals they could score along the way. Dreaming big is certainly better for raising money/stock price.

[jacob]

What always makes me curious is that no one solved the self-driving train problem yet (apart from trivial cases).

Do you mean a big train or a light train?

My wild guess is that driving the train is easy and that the reason for the existence of the engineer is to handle exceptions with big consequences.

[zbigi]

Do you mean a big train or a light train?

My wild guess is that driving the train is easy and that the reason for the existence of the engineer is to handle exceptions with big consequences.

Big trains. Light rail has a bunch of automated success cases deployed already.

I agree that it's partly about edge cases. Bo edge cases with high consequences are also one of the main problems in current FSD solutions. Consequences are more dire with train crashes but, with cars there's risk of human death, which, as far as companies go, is dire enough.
I think it's mainly about the money. Everybody in the field knows that universal driver is super hard, pie in the sky project that may take decades. The way to get investors riled up about it is to promise to "disrupt cars" (market worth 100b a year), not to make train trainsport incrementally more efficient.

Btw here's a super-cool presentation by John Carmack, who has taken a plunge into AI a couple years ago and is now showing some of his results. It gives a good taste of how uber-frustrating practical robotics problems can be
https://www.youtube.com/watch?v=3pdlTMdo7pY

[Sclass]

Dreaming big is certainly better for raising money/stock price.

This reminded me of something. This is 100% correct. Uber did something like this pre ipo when their potential cab fare revenue couldn’t justify their volume of private investment in 2012. Fares weren’t a big enough market so they started a self drive project. It didn’t go that far.