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

[RichardL]

All the usual subjects (GJ, toilet boy etc.) are posting on X that Tesla have been forced to implement Baidu’s autonomy solution in a huge defeat. No mention that is appears to be using the maps only.

 

[Linkmeup]

I've been busy today, but I just want to say: WE'RE GETTING CARS THAT DRIVE THEMSELVES (AND BETTER THAN HUMANS). NOT SOMETIME IN OUR LIFETIMES. NOW. ACROSS MANY COUNTRIES.

That is all.

 

[Ktowntslafan]

I was given a time out and received a bunch of these messages (below). Apologies to mods and anyone else I may have offended. I’ve been to hell and back with the local “elite” trying to take me down, cause I rep Elon/Tesla hard here and they hate him. I hope everyone is enjoying the price action today. Cheers.

 

[alexxs88]

What is your assumption of dealer distribution costs? Unless you have actual data for all elements fo dealer distribution to compare with direct distribution I doubt your statement.
Clue: margin between MSRP and actual sales price is not normally a significant source of dealer profit nor OEM distribution costs.

it was a major shock to me the first time I saw such data. Since I'ev actually seen to for two EU OEM, one Japanese and two US. None were below 20% of total cost of goods sold. Tesla will, of course have a very different structure and exact parallel is not possible.

As I mentioned before, a good friend of mine was essentially a VP of the financing arm of BMW in my country. I can't remember right now (I asked him and posted last week), but it was somewhere around 15% iirc. This might be a difference between regions. Over here there's no way in which a car gets sold over MSRP by a dealer. So what I'm talking about is this: customer buys BMW from dealer for $60k, dealer buys it from the national importer who buys it from BMW for $50k. So BMW would've sold that car direct, their gross margin would've looked a hell of a lot better. Hope that clears what I meant and really curious to get more insight into your data.

 

[Stretch2727]

There are numerous types of risk for which loss severity and loss frequency are directly correlated. Examples include most classes of storm risk, where high frequency and high severity are directly correlated. hence, just try to buy storm risk in South Florida.

In automotive it varies by geography. Other things remaining equal major urban areas have high frequency of accidents but slightly lower severity.

Highly automated systems are different. Well-designed automation invariably reduces loss frequency. That applies to robotic surgery, driver automation, aircraft automation and many more too numerous to name. Just think about any arena in which well designed automation si deployed. Every time human error decrease loss frequency decreases, often by major proportions.

Take robotic surgery as an example, in major surgery, say, removing a bladder full of tumors and replacing with a new-bladder from the patients own large intestine. That, even a decade ago was and eight hour marathon with surgeons in shifts. Now the surgery lasts half as long, is far more precise and far less stressful for everyone concerned. Were loss severity not a factor we'd assume surgeon liability insurance would plummet in cost. Not so fast, if a robot makes a mistake...then there is a huge predisposition to blame the machine and the imprudent people who deployed it. Now there are entirely new liability targets and vastly more consequential loss severity. In the case, I am told reliability, all that new gear has reduced error rates by huge promotions but increase the loss severity by a major factor. Thus, pretty much a rate wash, I am told.

Similarly, commercial aircraft insurance and shipping insurance have had major rises in rates, not because of frequency but precisely because they are historically such attractive low risk. A couple B737Max crashes and a ship destroying the Key bridge and suddenly there are massive claims upon massive claims, Those were only three events. Three. They are coupled with ships being hijacked in the Red Sea, tornados, floods, hurricanes and earthquakes. Suddenly nearly every insurance risk is rising, despite quite low frequencies.

The problems are that loss severity when it comes is incalculably high. Traditionally reinsurers take what are called 'excess' policies at cheap rates. But, all those unrelated things suddenly have happened so reinsurance markets are severely impaired (this is a very long discussion, I'll skip it except to say Lloyd's).

Now we have a device that demonstrably reduces risk of accidents and injury. It's new but seems magical (rather like robotic surgery, in a way) so we expect rates to drop. Nope, not anytime soon. When the inevitable accident happens and somebody is killed, what happens. Just like in robotic surgery, suddenly there are new wealthily pockets to pick and compelling stories about being killed by an infernal machine.
It takes little imagination to understand, then, that loss severity rises.

That is the way it works. Nothing arcane. It's really simple. The more dramatic effect of the loss before a jury, the higher the loss severity. Therefore automation proven benefits are indisputable, but when a problem happens...drama ensues. Actuaries HATE drama!

Ok, but 60% of my current auto policy is for the collision and property damage. Wouldn't this risk be greatly reduced with FSD as most of this is non-injury fender benders.

I think you are saying the liability will rise for accidents with major injury and death?

 

[SOULPEDL]

That is the way it works. Nothing arcane. It's really simple. The more dramatic effect of the loss before a jury, the higher the loss severity. Therefore automation proven benefits are indisputable, but when a problem happens...drama ensues. Actuaries HATE drama!

Do you mean like when those kids ran a Tesla into a tree and burned to death? Until we heard of that Taycan in a garage, plus many more warnings to park other EVs away from buildings.

I do see your point when Tesla Automation/Insurance is 100% responsible for an accident. That's freakin Terminator stuff. Maybe even Ride Insurance costs more than the ride - knowing how I keep hearing Learner and Rowe injury lawyers all day in Az. It's a leap of wills to cross this.

Your heed is certainly worth considering so as not to bet the farm and consider TSLA swings are still likely, but we gotta plow through this phase. Perhaps why FSD may not deploy in USA first, or in the way/location we're all assuming. I'd imagine there isn't the same level of abuse overseas. Some might welcome safety and cost savings in another land. And maybe it's all about the maps and how the Chinese really do want to save their current population of children.

 

[alexxs88]

1. I already told you "when". I said when people realize that driving manually is too dangerous. Insurance companies will give huge discounts for using FSD. That alone will drive the take rate high.

2. Yes, gross margins above 80%. Rather than give you one link, just google "profit margin for SaaS companies". I will say that the margin on FSD is hard to compare to anything else. It requires a lot less development and support personnel, but it does require a lot of expensive compute. In Tesla's case the compute resources will be shared with many non-FSD projects. So it's really hard to tell what the margin will be. My experience in software is that when you scale up and your software gets really popular, 90% margins are common.

3. OEM Deals are realistic today. As in NOW. Elon already talked about it on the conference call.

1. Why isn't this happening already? Tesla has been publishing data for years. More than anything, why isn't Tesla insurance orders of magnitude cheaper than other insurers if the cars are that much safer?
2. I'll give you 85%. Heck, I'll give you 99%. It doesn't change the maths. 6m (their current run rate is 3m) cars x 12 x 10 x 100$ x 30% ~= 22B profit. And this is beyond optimistic, borderline delusional because it ignores: discount future cashflow to present value, assume take rates an order of magnitude higher than today, assume someone buying a $15k 9-year old Tesla is happy to pay 10% of its value for a year of FSD. And so on...
3. Anything is realistic today. A good investment exercise is to wonder why isn't it happening already? And what will be the catalyst that will make it happen in the future and how likely is it.

 

[Dutchie]

You really shouldn't... his vids are arguably the least honest on the topic of all the influencer testers.

He was posting videos claiming FSD drove him to his favorite restaurant without interventions before fsdb was even a thing.

As he often does the video was greatly sped up.... when you slowed it down you could clearly see at every intersection he was manually making the turn then rengaging since FSD at that time didn't DO turning on city streets.

When someone in the comments pointed it out he said he meant any "unexpected" interventions, and manually making turns was expected, so he stood by the video title.


If you want a vastly more realistic version of where FSD is stick to folks like Chuck Cook.... who just recently posted his famous UPL testing on 12.3.6 and it's... not good....especially for a situation where Tesla had been directly training on that exact intersection for months and months now.

Chuck always makes these videos right after he installed them. My experience is that it takes a couple of days or weeks for a new version to “settle in”.

In the beginning FSD it always performs way worse than after a couple of days.

 

[anthonyj]

how much did that roof cost you? I was quoted $185k (lol) last year for a roof in the panhandle.

170k for 25.3kw with 3 powerwall 3s. $120k after tax credit. I figure a new roof is like $70k anyway, so an extra $50k for solar and batteries is a very fair price. I paid for it with a little swing trade I did last year, bought at $130ish and sold at $299 day of earnings (damn I’m good at this).

 

[alexxs88]

As time goes on, the timeline for the FSD singularity gets compressed. Thus, someone who is optimistic today in this regard will be much more likely to be correct than someone who was optimistic 3-4-5 years ago. After periods where pessimism wins out (recent 3 years let's say), extreme optimism is bound to rise to the occasion. This is how technological adoption works when rising out of the Trough of Disillusionment. Tell you what, let's talk Spring 2027 (3 years form now, if you are still around then) and see where we are at. We will see how the OEMs have fared; and if Waymo (as you suggest) is the market leader with OEMs adopting their hardware/software suite or if, rather, Tesla has chosen the correct path. Place your bets...

I don't think technological advancements like the ones required to reach autonomy work like that. Researchers have been battling various forms of cancer for centuries, each type had its "FSD V12.3" moment(s). Until it's solved, it's not. No matter how close it looks to the untrained or uninformed eyes from the outside.

But yes, happy to come back in 3 years time and see where we are then. For the record, I don't think we'll be in a significantly different position fundamentally.

 

[Stretch2727]

I've been busy today, but I just want to say: WE'RE GETTING CARS THAT DRIVE THEMSELVES (AND BETTER THAN HUMANS). NOT SOMETIME IN OUR LIFETIMES. NOW. ACROSS MANY COUNTRIES.

That is all.

If you drive where I drive, better than a human is not that difficult.

 

[Ktowntslafan]

I was given a time out and received a bunch of these messages (below). Apologies to mods and anyone else I may have offended. I’ve been to hell and back with the local “elite” trying to take me down, cause I rep Elon/Tesla hard here and they hate him. I hope everyone is enjoying the price action today. Cheers.

View attachment 1042924

I went with a new Avatar. Love ya'll. When do we break $200? This week?

 

[Usain]

1. Why isn't this happening already? Tesla has been publishing data for years. More than anything, why isn't Tesla insurance orders of magnitude cheaper than other insurers if the cars are that much safer?
2. I'll give you 85%. Heck, I'll give you 99%. It doesn't change the maths. 6m (their current run rate is 3m) cars x 12 x 10 x 100$ x 30% ~= 22B profit. And this is beyond optimistic, borderline delusional because it ignores: discount future cashflow to present value, assume take rates an order of magnitude higher than today, assume someone buying a $15k 9-year old Tesla is happy to pay 10% of its value for a year of FSD. And so on...
3. Anything is realistic today. A good investment exercise is to wonder why isn't it happening already? And what will be the catalyst that will make it happen in the future and how likely is it.

The catalyst for all 3 is end-to-end FSD. It just started happening. Have you not been paying attention?

 

[GeoX750]

Insurance companies will offer a huge discount for miles driven on FSD. If supervised FSD stays at $100/month, a lot of customers will save enough on insurance to justify the price.

When FSD works, Tesla will be liable for any accidents (as I will be reading a book or asleep). We will save a ton of $$$ as we shouldn't need liability insurance at all. Just like I don't need liability insurance to ride the bus or take a taxi.

 

[unk45]

As I mentioned before, a good friend of mine was essentially a VP of the financing arm of BMW in my country. I can't remember right now (I asked him and posted last week), but it was somewhere around 15% iirc. This might be a difference between regions. Over here there's no way in which a car gets sold over MSRP by a dealer. So what I'm talking about is this: customer buys BMW from dealer for $60k, dealer buys it from the national importer who buys it from BMW for $50k. So BMW would've sold that car direct, their gross margin would've looked a hell of a lot better. Hope that clears what I meant and really curious to get more insight into your data.

The major dealer items for margin tend to be service and warranty repair, excluding collision and used vehicles as well as Finance and insurance. All vary greatly by country.

 

[alexxs88]

The major dealer items for margin tend to be service and warranty repair, excluding collision and used vehicles as well as Finance and insurance. All vary greatly by country.

Are you saying that in the US, the dealer pays MRSP for the car? (or close to it) Respectfully, I feel like you're evading the answer. If BMW in Atlanta sells a 5-series for MSRP (let's say $60k), how much would they have paid BMW for that car?

 

[2daMoon]

Well, only up 35% over the past week.

I hope things pick up, as I'm feeling despondant, suffering through this dreadfully slow uptick.

 

[alexxs88]

The catalyst for all 3 is end-to-end FSD. It just started happening. Have you not been paying attention?

It's been happening for 8 years for Tesla. People have been working on AI for decades. Until we reach 100%, there's no way of knowing if we're 9% there or 90% there as a timeline.

 

[heltok]

What strikes me most is how well Elon is received by China and how sad it is that we in the west are not cheering for our guys anymore. Even if you disagree with Elon on some issues, which most Chinese do, we want the innovation to be happening here...

https://twitter.com/x/status/1784835703743881256

Europe has lost it a long time ago. America is still pushing ahead, but the administration and media sure could do better at cheering for the home team even if some of their players write tweets they disagree with.

As for Baidu I think this will mainly be Tesla buying maps for navigation. But it could be that Tesla will use Baidu infrastructure to store their video data and maybe to buy compute:

Baidu Cloud Compute (BCC)-Baidu AI Cloud

Baidu Cloud Compute (BCC)

intl.cloud.baidu.com

I think what will happen is that Tesla will iterate and train many neural networks in USA. This requires tons of compute thus they are expanding compute. Once they have some well trained neural networks, they need to refine these for China. Basically they will take the already trained neural network, then train it some more with a mix mostly new data and a little bit of old data. The amount of compute needed for this is large but a magnitude less than what they do in the USA. So they can probably afford to just purchase this compute outright. And Baidu is selling compute.

Tesla would need to ensure that their proprietary algorithms for training FSD V12+ is not stolen though.

I give it a 10-25% probability that Tesla is buying compute from Baidu.

 

[Artful Dodger]

I went with a new Avatar. Love ya'll. When do we break $200? This week?

Haha, that explains the K-Town in Kingston. Were you really roommates with Elon, that's quite a picture. Cheers!