Tesla, TSLA & the Investment World: the Perpetual Investors' Roundtable

[ammulder]

my general idea is that we have a massive conservative bias with our Tesla predictions.

Well, there’s also the minor issue that FAANG and friends have grown up and become mature and wildly successful and are valued at 1-2T.

Tesla has barely gotten started in terms of their eventual production, in vehicles, batteries, solar roofs, FSD, and etc. It seems quite presumptuous, if not a little crazy, to say Tesla at this early stage should be nearing the valuation of all those mature and successful tech superstars.

Not that I mind, exactly...

 

[StealthP3D]

Tried to warn you and others- My rule is not to sell covered calls when the stock is climbing steady.

Also, it's a good rule to not sell any covered calls before any dramatic rises that are completely unexpected.

/s

 

[MC3OZ]

There are many sizes and configurations of child seats. Logistical nightmare.

4-5 combinations cover most of the market, however there ae detachable capsules where a sleeping child can be removed from the car, simply by removing and carrying the capsule. Logistically, if the parents want to remove the capsule, then the Robo-taxi needs to wait for their return journey.

There is also a lot of gear parents with very young children need to carry around.

Once the youngest child in the family is say 2-3 years old, it is a lot easier.

Say there were 5 combinations of seats and 3 positions where children could be seated, the app to book the Robo-taxi would need to cover all options.. Most probably the family would have a profile which retained information on seating requirements, when booking a Robo-taxi the family would need to nominate the individuals that are travelling.

Computers can do all the matching and a certain portion of the fleet can be set up with the specialist seating configurations, however a family would need to book well in advance, or perhaps wait for a vehicle with their seating configuration to become available

Another option is quick docking car seats that can be installed very quickly in the Robo-taxi, a depot could install these as needed to help achieve a better match between supply and demand, that imposes extra costs, but a family with young children will pay for the extra cost if it is cheaper than owning a car.

An even bigger issue is cleaning the car seats, given the tendency of very young children to have accidents of all kinds. That suggests removing seats and a automated cleaning process.
If the seats need to be removed and cleaned after every trip, that is an extra cost.
Seem to me, at a minimum, they need to be quickly inspected.

Maybe they have seat covers which the family/depot must install and remove...IMO most likely the depot.

Given the logistical complexity, I can't see this being available on day 1. It is a tricky problem, not easy to solve efficiently.

 

[ReddyLeaf]

Mostly agree, but there's one class of people that will resist robotaxi, and that's parents. Hauling the child seat in an out of a robotaxi is a real pain--and what do you do with it while your shopping? Robotaxi will only be a success if it's also more convenient. If it's not more convenient (say you have to wait fifteen minutes for the robotaxi to arrive--five minutes is about the maximum) just lower cost alone won't drive the numbers you are suggesting.

To really make this work Tesla owners have to be convinced to use their car as a robotaxi. That would put enough cars in the area to get within the five minute time. Certainly robotaxi will eat normal taxis and Uber type services, and people where the cost sells them (of course, these are pretty large numbers, but nowhere near 50% as the taxi and Uber are often people who don't own a car to start with).

Something that I hadn’t thought about until your post, Tesla robotaxis will probably us AI, like their Autobidder. Therefore, the taxi fleet will always be properly spaced and located to minimize travel time and distance to the next predicted clients. Perhaps, the fleet will be so efficient that most pick ups will be nearly instantaneous. With camera vision, maybe one will even slow down, and wait for that stupid, slow human to finish typing the request into the app.

 

[Carl Raymond]

I find that, in the recent climate, there is a belief that rich ppl tends to side with conservatives. And they tend to have a world view where getting rich is a zero sums game. Cause to get there, you basically have to be an asshole.

Therefore most rich ppl are evil.

What you are observing here are ppl who managed to get there under the condition where others making it does not mean you have to lose it. A non zero sum game condition.

In old world conditions, it is alot harder to get this state of mind, but with internet that allows for group participation, getting rich doesn't have to be a zero sum game. The more ppl understand this, the less conflict there will be in the world.

I’m not convinced that’s entirely true. We often hear the expression “Tesla will eat their lunch”. We are getting rich off the back of Tesla.

Granted, if total GDP increases, it’s possible for there to be more wealth in the world, and with circular flows of resources and greater efficiency, Tesla does strive for that. But I’m under no illusion that nobody is getter poorer as we reap the rewards of picking the winner - short sellers for one.

 

[JustMe]

So here’s how it’s going to go. You’re not getting t

But I’m under no illusion that nobody is getter poorer as we reap the rewards of picking the winner - short sellers for one.

The very thought grieves me!

 

[LTC_RRR]

There are many sizes and configurations of child seats. Logistical nightmare.

Who gets to clean up the poop, the pee, and the puke after the ride?

 

[Criscmt]

That would be most awesome... cancel it now only to be forced to include it in a few years at 10x the current valuation.
They missed their best train, all subsequent ones will be worse and worse.

Inclusion is not bad for S&P, non-inclusion is. And is also getting progressively worse.

I agree that non-inclusion is bad.
I wonder about the possibility of gradual ramp in weightage over several quarters. Indexes won’t necessarily get at a lower price as the SP likely will keep going up over those quarters likely slowly, but squeeze likely will be avoided.
Thoughts anyone?

 

[astrotoy]

I think your CAGR calculator's busted.

Your 2013 shares are at about 55.6% CAGR right now, and your Jan 2015 shares are at about 57.4% CAGR.

You can back-check your 1st example easily with Google using this simple formula:

$26*1.556^7 = $574 so CAGR is very close to 55.6%​

Cheers!

I was actually dividing the gain into two parts. One was from purchase to November 2019, and the second was from November 2019 to now. My mistake was the length of time from purchase to November 2019, really 6 years and 4.8 years. So my gain from purchase to November 2019 was something like 17% and 14%, not too bad. However, the past year has been a crazy ride since then. I tried to edit my post, by TMF didn't let me.

 

[PeterJA]

There are many sizes and configurations of child seats. Logistical nightmare.

A first-principles thinker would ask: What is a child seat for?
Answer: To safely restrain the child with a child-sized harness.

So maybe the solution is harnesses in every car that are adjustable to accommodate all sizes of children. Again, I believe the first-principles thinkers at Tesla will solve the problem when it reaches their priority list.

Who gets to clean up the poop, the pee, and the puke after the ride?

The parents, or they won't get to ride again, as I wrote awhile ago:

Tesla, TSLA & the Investment World: the 2019-2020 Investors' Roundtable

 

[amolina]

FSD Beta reaching Canada according to Twitter
https://twitter.com/FSDBetaCanada/status/1332050678047510531

 

[The Martian]

* Tesla's DoJo supercomputer and NN training farm has a bigger value than Amazon's AWS

Agreed. Among other points, the above point is even overlooked by a lot of Tesla bulls. In below tweet, Cobra Kai is referring to Dojo. I remember in other Dojo related Elon tweets, when someone asked if Dojo will be available to other people for NN training, Elon said Yes. That mean Tesla intend to open Dojo to public to become AI of AWS. Keep in mind that around 57% of Amazon's operating income came from AWS in Q3 2020. The potential of Dojo is beyond FSD and could be huge.

 

[BigPeePee]

Agreed. Among other points, the above point is even overlooked by a lot of Tesla bulls. In below tweet, Cobra Kai is referring to Dojo. I remember in other Dojo related Elon tweets, when someone asked if Dojo will be available to other people for NN training, Elon said Yes. That mean Tesla intend to open Dojo to public to become AI of AWS. Keep in mind that around 57% of Amazon's operating income came from AWS in Q3 2020. The potential of Dojo is beyond FSD and could be huge.

View attachment 612437

Wouldn't that help competitors who are building FSD?

 

[Beltsbear]

Agreed. Among other points, the above point is even overlooked by a lot of Tesla bulls. In below tweet, Cobra Kai is referring to Dojo. I remember in other Dojo related Elon tweets, when someone asked if Dojo will be available to other people for NN training, Elon said Yes. That mean Tesla intend to open Dojo to public to become AI of AWS. Keep in mind that around 57% of Amazon's operating income came from AWS in Q3 2020. The potential of Dojo is beyond autonomy driving and could be huge.

View attachment 612437

Nope. The revenue from capturing 100% of the NN training market even if it was 10x the size now would not be that much. Tesla built a very good neural chip and Dojo has the potential to be great, but others are doing well in this space too. Tesla will not get anything more then 50% of the NN trading market. Most neural network training can be done on available equipment and does not need Dojo.

I would think Tesla’s Dojo , battery and network building experience will make Tesla robotics a much bigger deal then Dojo alone.

 

[Big Time]

Supposedly Elon is now worth 142B compared to Bezos 187B. At what Tesla share prices does he pass him?

 

[BigPeePee]

Nope. The revenue from capturing 100% of the NN training market even if it was 10x the size now would not be that much. Tesla built a very good neural chip and Dojo has the potential to be great, but others are doing well in this space too. Tesla will not get anything more then 50% of the NN trading market. Most neural network training can be done on available equipment and does not need Dojo.

I would think Tesla’s Dojo , battery and network building experience will make Tesla robotics a much bigger deal then Dojo alone.

Tesla robotics?

 

[Johnny Ma]

I always considered myself a super bull on Tesla, but the stock has gone up considerably faster than myself and pretty much everyone else imagined. I have several thoughts around this topic, but my general idea is that we have a massive conservative bias with our Tesla predictions.

Take a look back at some of the old super-bull predictions, which now seem like bear predictions: Super Bulls Only

Or take a look at ARKs 7000EV (1400) by 2024 target... starting to look kind of low, especially with FSD looking much closer now than people thought.

I think the reason for having a conservative bias is:
* Unconscious absorption of FUD
* Expecting the past flat period from 200-300SP to repeat, instead of realising we are on an S curve of ramping
* Struggling to imagine the scale of Tesla/EVs/renewables
* Struggling to imagine the new tech, and the changes they bring
* Underestimating FSD and more specifically the effectiveness of computation scale on AI (May 2020 news & 'On GPT-3' · Gwern.net)
* Not wanting to sound crazy

I think the most accurate predictions of Tesla will be top down based on total addressable market, across all the markets they are touching, with massive margins, growth and market share. From EVs, to AWS for AI and all 15-20 or so industries in between. In this regard you can think of Tesla as 15-20 companies, all which can end up leading their fields. 550B might seem like a lot now, but really it's just 20 companies undervalued at 30B each. Some of those startups perhaps could be valued at 550B alone... most notably FSD.

It has been on my todo list to attempt sizing up these markets and assigning a large percentage of them to Tesla. For now though, I can say the potential upside is so much higher than I see anyone putting numbers to.

I would like to see an analysis where:
* Tesla gets to FSD in next 2 years with high probability and has 5 years with no competition in FSD
* Tesla batteries/energy/EVs grows 20% faster than predicted at battery day
* Solar scales at similarly rapid rates
* Tesla's DoJo supercomputer and NN training farm has a bigger value than Amazon's AWS
* Decade long predictions take into account acceleration change: Accelerating change - Wikipedia
* And comprehensive valuation of all the other startup components like insurance, chip design, seats, etc.

Taking all this into account, I can see a case for a current SP of 2000-3000. I may sound crazy now, but check back in 5-10 years and see whether I was crazy, or perhaps even bearish in my outlook. Perhaps even 1-2 years will prove me right. I should say I am not 100% confident on any of these super-bull predictions, they are complete guesses and I haven't done the top down analysis which I think is needed to most accurately predict the 10 year stock price. It's not really necessary for a buy and hold investor like me.

However, I did want to point out a historical collective bearish shortcoming in predicting Tesla and TSLA.

Been a while since I've visited but just needed to say:

Wow. Incredible post. Glad more people know who Gwern is.

Anyway, I'm not selling even after today's, the week's, the month's, the year's SP action. You need to think a lot bigger if you're going to get my shares.

 

[ggr]

Mod: Enough car seat stuff. And other OT stuff too. --ggr

 

[Beltsbear]

Tesla robotics?

A self driving car is a robot. The same kind of neural network structure amplified will probably power the first practical home robots. While I know of no work by Tesla in this field, Tesla is positioned to enter into this field in the future. And home robots could be the price of an expensive car. First sold to the rich.

 

[BigPeePee]

A self driving car is a robot. The same kind of neural network structure amplified will probably power the first practical home robots. While I know of no work by Tesla in this field, Tesla is positioned to enter into this field in the future. And home robots could be the price of an expensive car. First sold to the rich.

Okay, that makes more sense.