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

[FastEddieB]

I have never taken out an auto loan in, what, five decades of owning cars.

The above is, I'm sure, no surprise to anybody who hangs out on an investment forum. But it's amazing how few consumers realize how bad the rat race of buying stuff on time is.

Good post.

We usually pay cash and stay within our means. We’re fortunate enough to be able to.

But we will finance if interest rates are low enough. Just recently paid off a 60 month tractor loan at 0%. Financed a motorcycle at 1.99%. In these case, it can makes sense to keep the money working for you via investments, interest and dividends so long as you can earn at a rate in excess of the loan rate.

 

[Usain]

The pessimistic side: Elon has repeatedly said the closer we get the more it will cost....

Yes, that's what is so confounding. With a massive increase in training hardware, I expect that we will see rapid improvement in FSD very soon. And when that happens it seems that the price of FSD will have to increase dramatically. Or, Tesla switches to a subscription-only model.

The one thing I think we can say for sure is that the FSD price cut means Tesla wants to increase the take rate immediately. But to me, the motivation is unclear.

It could definitely be a bearish sign. But that just doesn't square with all the positive FSD news we have had lately. That's why I continue to believe that the FSD price cut has something to do with needing more FSD users on hardware 4.

 

[Usain]

FWIW I bought FSD for £10,000 ($12,000) (for the second time!). Nobody I know would pay that much for it right now. Most people I know wouldn't even pay that if it was complete autonomy.

If Tesla achieves complete autonomy, the numbers get crazy.

My aging parents would pay far more than $12k if FSD could reliably get them from place to place without intervention. They aren't comfortable taking Uber and they are really too old to be driving themselves.

How many people in America are unable to drive themselves but would be willing and able to pay $25k for a multi-year contract with a chauffeur that is available 24/7? If it's 10 million, that's $250 billion in pure profit before we even start to think about robotaxis.

(There are about 23 million people in the US aged 75 and over.)

 

[Snerruc]

I sell software. The key differentiator that makes software nothing like most businesses is that the marginal cost of production is zero.
FSD has to be priced at the point at which total revenue from FSD is maximised. Assuming FSD only has one flavour, that means the optimal pricing is such that a bunch of super heavy users who have high incomes are going to get an amazing deal, but thats just how pricing works...

It could be that the optimal price for FSD is $50,000, or it might be $2,000. Tesla can never know unless they vary the price and observe. People are VERY bad at predicting this stuff, and even predicting their own decisions at various price points until they are confronted with them.

FWIW I bought FSD for £10,000 ($12,000) (for the second time!). Nobody I know would pay that much for it right now. Most people I know wouldn't even pay that if it was complete autonomy.

Hmmmm…. I seem to remember that anti lock brakes were an expensive option on expensive cars when they first came out. I remember the same on back up cameras. Now they are standard on a Tercel for no up charge. Or any car. Governments tend to mandate proven safety features. Why would self driving be different? The window to make big bucks may be small.

 

[hobbes]

Fun anectode before market opens: Family of a friend just bought a Model 3 (my 4th sale ). They seem really happy with it, get messages all the time showing how low they can get the Wh/mi. Their son is in college becoming a sports teacher and just went to the french coast for a surfing workshop (made me realize I studied the wrong thing...). First day he hurt his knee so bad he has to go back :/. Just got a text from his mom, "on my way in Model 3 to Bordeaux" - turns out it´s not her driving to France but quoting him that a taxi is picking him up for the airport (organized by insurance) .. Just to illustrate how common 3s and Ys have become in Europe.

 

[Artful Dodger]

The one thing I think we can say for sure is that the FSD price cut means Tesla wants to increase the take rate immediately. But to me, the motivation is unclear.

I think they're paying for the price cuts on S/X with an increased take rate on FSD. For example, going from 10% adoption to 12.5% replaces the foregone margin on S/X cars. Plus, there's the bonus @cliff harris spoke about upthread: price discovery.

Maximizing profits means the area under the curve for unit price vs sales volume. Tesla can only know this empirically by varying FSD prices and watching the effect on FSD take rate. This is Management Science 101, hence is completely lost on Wall-E.

That's why I continue to believe that the FSD price cut has something to do with needing more FSD users on hardware 4.

I don't think that's a real concern going forward. I believe that all Tesla cars with HDW 3 or above will run the appropriate FSD stack even for basic autopilot, just with a reduced feature set. The data collection will stil occur as the rest of the FSD stack runs in shadow mode: that is watching the driver's actions and comparing it to what FSD would have done, and reporting any discrepancies via the normal WiFi uplink process used in FSD Beta.

Cheers!

 

[DragonWatch]

Forward Observing

Humor ~ as a teenager, McDonald’s announced the sale of their millionth burger. Someone started the joke at the time; well that means they just killed another cow. Hopefully you get it; you do not need to have been back in time to get it.

Full Self Driving (FSD)/Autopilot ~ has always been the pie in the sky. Therefore, the details have never been held in front of you/me or the other human animals. My introduction to FSD was years ago ~ simply known as Cruse Control (CC). Cruse control took years to introduce and develope. Today, cruse control is standard on the average fossil fuel car. Bottom Line to become absolutely implemented, every car on the road must have it; in the case of FSD. Otherwise, the fear of the unknown will rule. Refinements will always occur. At some point it will become standard ~ if not, it will never become standard option; and the goal in the pie in the sky is that everyone one/car, truck has it. One of the first stages is moving from Beta, then hands free, then the comfort zone, and then one day you will wake up talking to yourself in a small, presumably four wheeled box, called a car taking you off to wonderland.

I may or may not be there, but cool to be here now.

Cheers

 

[jkirkwood001]

By its comparative analysis of the following, Yahoo Finance suggests TSLA is undervalued and its fair market value should be $455.84 according to it's "GF Value":

1. Financial strength
2. Profitability & growth
3. ROIC vs WACC

Unveiling Tesla (TSLA)'s Value: Is It Really Priced Right? A Comprehensive Guide

Exploring the intrinsic value of Tesla (TSLA) using GuruFocus' proprietary GF Value

finance.yahoo.com

Call me sceptical, but prognosticators sometimes sound hollow. It's like someone saying "this team should have won the World Cup / Super Bowl / World Series", but they didn't, so what's your point? We're agreed that the stock market is maybe half logical but no more, so there's no point in arguing why investors haven't invested. They just haven't, point finale.

Okay okay, despite it all, I still believe in TSLA's growth and am in for the long haul! HODL.

 

[GOVA]

Any news/rumors if/when MY will get Highland upgrade? Seems like a natural progress once M3 has been upgraded.

It is not bad now, but with new suspension and all other gizmos, MY would be perfect.

 

[mongo]

Yes, that's what is so confounding. With a massive increase in training hardware, I expect that we will see rapid improvement in FSD very soon. And when that happens it seems that the price of FSD will have to increase dramatically. Or, Tesla switches to a subscription-only model.

The one thing I think we can say for sure is that the FSD price cut means Tesla wants to increase the take rate immediately. But to me, the motivation is unclear.

It could definitely be a bearish sign. But that just doesn't square with all the positive FSD news we have had lately. That's why I continue to believe that the FSD price cut has something to do with needing more FSD users on hardware 4.

Tesla doesn't need people to buy FSD to leverage the car and hardware for training data.

FSD price always needed to include the contingency cost of a hardware upgrade. If they are confident in that no longer being a risk, they can reduce or remove that amount from the price (a form of warranty reserves).

It also covered the R&D costs of development. If the labor and compute costs (present and forward looking) are reduced along with an increase in sales, it reduces needed pricing.

 

[Usain]

Hmmmm…. I seem to remember that anti lock brakes were an expensive option on expensive cars when they first came out. I remember the same on back up cameras. Now they are standard on a Tercel for no up charge. Or any car. Governments tend to mandate proven safety features. Why would self driving be different? The window to make big bucks may be small.

If the government mandated self driving then the numbers get even crazier for Tesla. Tesla would be the only vendor and every OEM would have to license Tesla's FSD. So just in the US, that's about 14 million cars. And say Tesla "only" charges $10,000. That's $140 billion per year.

 

[Buckminster]

Threads of the day:

Any news/rumors if/when MY will get Highland upgrade? Seems like a natural progress once M3 has been upgraded.

It is not bad now, but with new suspension and all other gizmos, MY would be perfect.

Maybe a year or so:
MY 2024 Refresh - Project Juniper?

 

[FreqFlyer]

Any news/rumors if/when MY will get Highland upgrade? Seems like a natural progress once M3 has been upgraded.

It is not bad now, but with new suspension and all other gizmos, MY would be perfect.

Last I read was end of next year for the Juniper MY.

Highland M3 at the beginning of next year, will be there be a Tesla debut event? Who knows.

 

[ItsNotAboutTheMoney]

If the government mandated self driving then the numbers get even crazier for Tesla. Tesla would be the only vendor and every OEM would have to license Tesla's FSD. So just in the US, that's about 14 million cars. And say Tesla "only" charges $10,000. That's $140 billion per year.

Now you're getting into fantasy territory. The more good AVs there are, the _less_ reason there is to mandate it. Much more likely would be governments getting tougher on driver testing and OUI.

In addition, people forget that there's ADAS competition as well.

 

[Usain]

I don't think that's a real concern going forward. I believe that all Tesla cars with HDW 3 or above will run the appropriate FSD stack even for basic autopilot, just with a reduced feature set. The data collection will stil occur as the rest of the FSD stack runs in shadow mode: that is watching the driver's actions and comparing it to what FSD would have done, and reporting any discrepancies via the normal WiFi uplink process used in FSD Beta.

Cheers!

I don't think shadow mode is enough. Tesla will want real disengagement data from hardware 4 FSD users and they will want lots of HW4 users as soon as they can get them.

 

[Skryll]

One weird data point: Hong Kong still has the old Model 3 (Est. delivery: Q3 2023). Hong Kong is right hand drive like Australia. I had always assumed HK is a harbinger to what we will see in AU (They got the Model Y before us).

Design Your Model 3 | Tesla

Design and order your Tesla Model 3, the car of the future. Learn about lease, loan and cash payment options, warranties, electric vehicle incentives, gasoline savings and more.

www.tesla.com

So is Giga Shanghai still running a batch/line with the old Model 3?

Edit: Macau is also RHD, a Special Admin Region of China just like HK, but gets the new Model 3 !! I am trying to understand the logic here.

Maybe China going to turn the screw on switching away from RHD /jk

 

[mongo]

I don't think shadow mode is enough. Tesla will want real disengagement data from hardware 4 FSD users and they will want lots of HW4 users as soon as they can get them.

Even if FSD isn't active, if the human does sonething different that the NN would have, then that can get flagged and submitted, just like a disengagement does (without voice note, of course).

 

[AudubonB]

There are some additional factors:-

• Vehicle Price reductions mean some hit to revenue and profits, even if the same percentage margins are retained.
• With the Cybertruck, Austin Model Y and Berlin Model Y ramps currently progressing slowly, there is no fast way of growing sales volumes.
• Shutdowns in China might also have a marginal impact on sales volumes.
• The higher FSD price was set before interest rate increases. (Rationale for dropping prices is similar to vehicles).
• Tesla might now have a much better idea about what they think it will cost to produce FSD.
• Perhaps the previous price was too high.

Now isn't a bad time to throw this demand lever, with apparent V12 progress, and the lower price, providing more of a buying incentive.

Any increase in the take rate will probably bring in significantly more Q3 revenue than would have been achieved by the higher price.

The price can also be increased again at a later date and advance knowledge of that increase used as a demand lever.

The percentage of the fleet who will have FSD at the end of Q3 will still be small in relation to the current fleet size, and very small in comparison to the likely future fleet size.

Those are good points, other than your first bullet.

Depending on where along the demand curve the relevant price points are, it is quite possible for both sales and profits to rise after a price cut, not fall. As Dodger wrote, you need to measure the area under the demand curve. Likewise, that could occur at the same time profit margins diminish.

 

[Usain]

Now you're getting into fantasy territory. The more good AVs there are, the _less_ reason there is to mandate it. Much more likely would be governments getting tougher on driver testing and OUI.

In addition, people forget that there's ADAS competition as well.

Yeah, I don't think it will be mandated. Someone was just suggesting that a mandate would drive down profits. I think it would have the opposite effect because Tesla would have a monopoly on the technology for quite some time.

I think the most likely way for someone to compete with FSD would be for Google to partner with an OEM like GM. The Chinese might be able to compete as well. But it would take a good 5 years for anyone to outfit a huge fleet with the hardware and curate enough data to build a comparable system.

Of course, all this assumes that Tesla's methodology is the way to go. More and more, this assumption seems to be correct.

 

[UltradoomY]

For those wanting Model 3 refresh in the U.S. (and Canada) don't hold your breath until Q4 and Cybertruck release event.

This will be the 1 more thing they add - sales and delivery will start immediately in Q4 with CT release.

Makes sense for financials and end of year push.

But hey Q4 is less than 25 days away.....