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

[Webeevdrivers]

I think Teslas driving on FSD should have magnetic stickers like those you see for student drivers, something like “Be Patient, AI Driver.”

We have this one on our car. We still get the occasional honk.

 

[unk45]

Tesla was going to build a plant with CATL but it wouldn’t have received the battery subsidy so instead it’s bought “spare” equipment from CATL to make the batteries.
The number produced will not be enough for a model 2.

I have not heard any other news about tesla building or acquiring battery’s from companies which will meet the subsidy criteria quickly.

So for this reason I suspect the robotaxi/ model 2 is many years away from mass production.

Tesla to Open US Battery Plant With Equipment From China’s CATL

Tesla Inc. will expand battery production in Nevada, opening a small facility using idle equipment from China’s Contemporary Amperex Technology Co. Ltd., according to people familiar with the matter.

www.bloomberg.com

This is plausible. CATL did that some years ago in brazil with Moura, to sidestep some onerous constraints at the time. To do such a deal with Tesla is entirely logical. The one major constraint is coming up with enough raw materials to preserve the US compliance.

 

[2daMoon]

Supervised accident rate has no strong correlation on unsupervised accident rate. Critical disengagement rate is a better proxy.

As I stated the absolute figures don't matter in my model as all parameters scale with miles driven. It's about cost / revenue ratios. There is no "breakeven" in the thousands of miles per critical disengagement rate, that's simply way too low as the costs of an accident are simply too high.

Um, yeah, I don't follow.

As the capabilities of FSD grow, the collision rate will go down. (it is already very, very low, correct?) So, it seems imperative that figures used to calculate the differences achieved for each version be compared in order to show progress in that regard. This is how to track how many miles will be driven before "X" happens.

These don't have to be absolute figures. But they can be broken down into classes, like deaths, injuries, damage to a RoboTaxi, damage to other's property, or even how many times per miles did the car pull over and park safely because it could not continue on its own. These are similar to the statistics that have been calculated for drivers for a very long time. Now, they can be calculated for autonomous cars.

There is also fault to consider when gathering such data. Particularly if the majority of collisions are the fault of non-autonomous (meat-sack piloted) vehicles. As more autonomous vehicles are put on the road, there will be an increasing ratio between them and meat-sack operated vehicles. Expectations are that one RT might replace three to five standard cars completely. This would need to be accounted for over time as well, as this is a dynamic system.

The way I see you using the term Critical Disengagement, it doesn't seem all that useful. What the autonomous vehicle did when facing a problem it cannot handle is important. I expect it to do something, not just close its cameras and go na na na na na na until the collision is over.

Maybe I don't understand your use of the term "Critical Disengagement" in regard to a vehicle operating solely on FSD with no option for a meat-sack in the vehicle to take over. If you could define this term in regard to FSD, with examples of a few minor to major "critical disengagements" that FSD might experience, this might be helpful.

The term "accident" is one I've avoided using in this context for years, though it is common parlance, it is not as descriptive as "collision" and I would like to verify this is how you are using the term. Do you mean collision when you write accident?

 

[Usain]

Well, I took the "extinctionist" comment as sarcasm, to emphasize his point. Though he may actually see where people might die unnecessarily due to some path someone has charted.

With this possibility in mind, he may be frustrated with people taking action who haven't worked out how the many challenges have to come together, carefully, over time, to culminate in reaching the goal with the least amount of negative impact to humanity.

Such negative impact might come from making rash decisions/legislation/etc. or, by getting gung-ho on accomplishing some part before another key thing has been achieved which that part's success depends upon.

Maybe, from his point of view he can see how things need to be strategically orchestrated. Like, we can't just turn off the fossil fuel tap while we still need it.

This statement may be him wanting to employ saving the planet tactics so as to avoid taking two steps forward only to have to take one step backward before continuing.

You may be right. I hope Elon doesn't really think that environmentalism is morphing into a human extionctionist movement. That's just silly.

But I could see how he might be mad enough at the people who knocked out the power in Berlin to ascribe almost anything to them.

 

[Musskiah]

Tesla was going to build a plant with CATL but it wouldn’t have received the battery subsidy so instead it’s bought “spare” equipment from CATL to make the batteries.
The number produced will not be enough for a model 2.

I have not heard any other news about tesla building or acquiring battery’s from companies which will meet the subsidy criteria quickly.

So for this reason I suspect the robotaxi/ model 2 is many years away from mass production.

Tesla to Open US Battery Plant With Equipment From China’s CATL

Tesla Inc. will expand battery production in Nevada, opening a small facility using idle equipment from China’s Contemporary Amperex Technology Co. Ltd., according to people familiar with the matter.

www.bloomberg.com

You are correct, the CATL equipment in Nevada will produce batteries to supply both Semi and Megapack 2.0, however...

Panasonic has a plant in Nevada, is building a second one in Kansas, and will announce a third one this year. It is likely they will be the supplier to Tesla for RT/M2. Panasonic says plans for third U.S. EV battery plant still on

"...the company, which supplies Tesla, still needs "more plants" to reach its goal of increasing annual production capacity from 50 gigawatt-hours currently to 200 gigawatt-hours by 2031."

Tesla-battery maker Panasonic may add 4680 battery capacity to Kansas plant

Tesla battery supplier Panasonic Energy is considering pouring new investments potentially up to $4 billion in its De Soto, Kansas, plant to produce 4680 battery cells for the Tesla Cybertruck and Tesla’s next-gen vehicles.

electrek.co

"Tesla battery supplier Panasonic Energy is considering pouring new investments potentially up to $4 billion in its De Soto, Kansas, plant to produce 4680 battery cells for the Tesla Cybertruck and Tesla’s next-gen vehicles."

 

[Musskiah]

Supervised accident rate has no strong correlation on unsupervised accident rate. Critical disengagement rate is a better proxy.

As I stated the absolute figures don't matter in my model as all parameters scale with miles driven. It's about cost / revenue ratios. There is no "breakeven" in the thousands of miles per critical disengagement rate, that's simply way too low as the costs of an accident are simply too high.










Costs could definitely be too low. I WAGed based on California, where electricity costs alone will end up being like $0.1 per mile. Then you have to amortize car purchase, among other expenses. Other states will be cheaper in some sense. I don't think the change in operating costs will affect the analysis too much.

Accident costs could be much higher, that's why I looked at the higher ($10,000 per accident) as well.




The TLDR is yes, cost won't be the impeding factor, but what you said likely will. It will still have to get near 100k / critical disengagment for public to accept it.

FSDBeta has run over 1B miles supervised without a fatality, most of these miles were less capable versions prior to V12. A Critical Disengagent does not imply an accident/collision, it just means FSD relinquished control to a human. FSD-S (supervised) V12.3.3 in it's current form has likely already drastically reduced, and in the very near future could eliminate the accidental collisions of Teslas (and other FSD licensing vehicles) with bicyclists and pedestrians, especially at lower speeds, say under 35mph. This alone is reason to demand the technology be used as often as possible. Do you know someone who died because of a bike or pedestrian accident? JB Straubel does....his wife! She died right around the time he left Tesla. Imagine how disappointing this was to him knowing that Tesla had technology that likely would have saved her life had it been in the vehicle that struck her.

Supervised, FSD is already far safer than any one of us driving alone. The data already showed this to be true of Autopilot all the way back deep into last decade. It is not difficult to reason that, at current rates of progress, FSD Unsupervised will become safer than most humans. Sure it may get confused and pull over, or cause drivers to become frustrated, but FSDBeta has shown fatalities have been all but eliminated ALREADY! Over 1B miles, without a fatality. By August 8, we are likely to see that metric get twice as impressive. This is not irrelevant data, to the contrary, it is very encouraging.

 

[AndrewZ]

Other than a random guess at a price target, once, that was still wrong on target date by years, and wrong on the specifics of how it got to that target, ARKs models are consistently, hilariously, wrong about nearly everything.

Why do we need to keep reminding people of this?


With tons of awful analysts throwing out targets all over the place SOMEONE was bound to accidentally "predict" a high price for Tesla-- Cathie was lucky enough for it to be her--- but it would've been a terrible idea to give her your money to manage as her actual results, over years, make clear.

Their financial performance has been awful- worse than just buying an S&P500 index fund, over a period where their one "win" (Tesla) has massively beaten the S&P500 over the same number of years.... because everything else they've invested in has done horribly.

They got lucky, once, picking Tesla early on- and just about everything else they've done since has been terrible.

Back in 2021, the bear case for next year was 5million sales and the bull case was 10million

Their bull case for three years from now is north of 20million sales

 

[M|S|M MYP]

Just took the dog for a walk and saw some solar contractors installing 2 powerwalls at a neighbor's house 2 doors down. FWIW, this guy owns 2 Fords, a F-150 and a High HP Mustang. Stopped by and talked with him and he said he's purchasing a Model 3 soon.

He said he did a year worth of research and nothing came close to the Tesla Powerwalls and he got a pretty sweet deal since we're here in SoCal.

I think its about time for me to check how much solar + Powerwalls are gonna run me

 

[Zaddy Daddy]

Um, yeah, I don't follow.

As the capabilities of FSD grow, the collision rate will go down. (it is already very, very low, correct?) So, it seems imperative that figures used to calculate the differences achieved for each version be compared in order to show progress in that regard. This is how to track how many miles will be driven before "X" happens.

These don't have to be absolute figures. But they can be broken down into classes, like deaths, injuries, damage to a RoboTaxi, damage to other's property, or even how many times per miles did the car pull over and park safely because it could not continue on its own. These are similar to the statistics that have been calculated for drivers for a very long time. Now, they can be calculated for autonomous cars.

There is also fault to consider when gathering such data. Particularly if the majority of collisions are the fault of non-autonomous (meat-sack piloted) vehicles. As more autonomous vehicles are put on the road, there will be an increasing ratio between them and meat-sack operated vehicles. Expectations are that one RT might replace three to five standard cars completely. This would need to be accounted for over time as well, as this is a dynamic system.

The way I see you using the term Critical Disengagement, it doesn't seem all that useful. What the autonomous vehicle did when facing a problem it cannot handle is important. I expect it to do something, not just close its cameras and go na na na na na na until the collision is over.

Maybe I don't understand your use of the term "Critical Disengagement" in regard to a vehicle operating solely on FSD with no option for a meat-sack in the vehicle to take over. If you could define this term in regard to FSD, with examples of a few minor to major "critical disengagements" that FSD might experience, this might be helpful.

The term "accident" is one I've avoided using in this context for years, though it is common parlance, it is not as descriptive as "collision" and I would like to verify this is how you are using the term. Do you mean collision when you write accident?

Um, the use to the metric of disengagements is common in autonomous driving evaluation. It's a metric that has to be reported to the California DMV, I believe, for instance.

A critical disengagement happens when the AV is about to hit something, and the driver intervenes. Certainly it is important in evaluating the competency of the system. You can have a system well operated that has 0 collisions, but is far away from ready for robotaxis.

Yes, both disengagements and collisions per mile driven should be recorded. If your system is truly capable of driving better than a human, then both will end up being superior to a human on a per mile basis.

 

[geometro]

Here is a hypothetical question: Could Model 3 rear-wheel drive price drop to the $25000-ish, and no need to create new "$25000" car?
I bought Model Y in Dec 2023, at $49990, and I even drove 82 miles to Stockton to pick up. And I had to file tax carefully to get $7500 credit.

Now since Q1 production is much higher than delivery, TESLA has to drop 3/Y price by $4000 to $5000 to clear inventory.
All-wheel Y is now $44990, real-wheel Y $39470. Rear-wheel 3 $36000.
If you deduct the Federal tax credit $7500, then Y is $31970 and 3 is $28500 (although 3 may not qualify).

Now the gap to $25000 seems worth of a stretch, especially since you can imagine how much performance and features could be cut-off in a $25000 car.

Moreover, if Gen 3 production process is really that good, production cost could be down and I believe another $4000 price drop is very possible.
In that case, why bother with the capital expense to develop and build a $25000 car production line? The only advantage for some buyers are now
just the smaller size of the car, if that is what they want.

That said, I am still worrying about TESLA delivery number for Q2. They must figure out a way to bring the number back to 450,000 level.

I don’t think it’s pricing at this point… the media narrative seems to have moved to EVs are a “bad idea” so I think more education and awareness is needed.
It wasn’t that long ago that Tesla was considered “cool” and aspirational… now it seems to be less so.

I told a co-worked that they are hands down the best cars to own…. Low ongoing cost, fun and no gas stations. Still believe this but he was definitely less interested than I expected.

 

[geometro]

I dont understand why there are concerns that there are insufficient batteries for a model 2? Its not like cheaper EVs than the model 3 have not been built by other companies. The MG4 exists, The Nissan Leaf Exists, there is a cheap peugeot EV. How are these other companies getting hold of batteries to make cheap compact EVs but Tesla cannot?
Also AFAIK Tesla are way ahead of any other car company when it comes to having control over their own battery supply chain, including giga Nevada, but also the 4680 line.
I know, I know... everyone is disappointed at the run-rate of 4680s right now, but we were all unhappy about FSD progress until recently. These things change fast!

As an investor, I'd like Tesla to announce and start making compact model 2s as soon as they can, EVEN if they are battery constrained to only make 100k a year. The fact that an affordable, small Tesla compact car exists will mean a lot of people make plans to get one as soon as the supply is available. Obviously from a mission POV, all EV sales are good, but as an investor I'd rather people put a reservation in for a M2 than an MG4 or Leaf .

Also, a small compact car is a new segment for Tesla. Not a bad idea to get a vehicle out there in customers hands, even in (relatively) small numbers so it can be perfected before it gets made in the millions?

I would not be surprised to see robotaxi AND model 2 announced and demonstrated on the 8th.

At least in the US, if I was renting an apartment I don’t think getting an EV would be high on my list. Charging would be too inconvenient and these may be the same demographic that the pricing of the Model 2 is trying to attract.

 

[Zaddy Daddy]

FSDBeta has run over 1B miles supervised without a fatality, most of these miles were less capable versions prior to V12. A Critical Disengagent does not imply an accident/collision, it just means FSD relinquished control to a human. FSD-S (supervised) V12.3.3 in it's current form has likely already drastically reduced, and in the very near future could eliminate the accidental collisions of Teslas (and other FSD licensing vehicles) with bicyclists and pedestrians, especially at lower speeds, say under 35mph. This alone is reason to demand the technology be used as often as possible. Do you know someone who died because of a bike or pedestrian accident? JB Straubel does....his wife! She died right around the time he left Tesla. Imagine how disappointing this was to him knowing that Tesla had technology that likely would have saved her life had it been in the vehicle that struck her.

Supervised, FSD is already far safer than any one of us driving alone. The data already showed this to be true of Autopilot all the way back deep into last decade. It is not difficult to reason that, at current rates of progress, FSD Unsupervised will become safer than most humans. Sure it may get confused and pull over, or cause drivers to become frustrated, but FSDBeta has shown fatalities have been all but eliminated ALREADY! Over 1B miles, without a fatality. By August 8, we are likely to see that metric get twice as impressive. This is not irrelevant data, to the contrary, it is very encouraging.

This is a nice strawman argument.

No one said Supervised FSD isn't useful for reducing accidents.

Its metrics don't tell us how close unsupervised FSD is to human performance.

The "current rate of progress" doesn't tell us exactly when it will exceed human performance. If you relied on the rate of progress from 2020-2023 then you would conclude it wouldn't reach robotaxi level for decades. Sounds silly right? The fact that miles / critical disengagements went from 100 to 400 in the last 6 months doesn't mean it's going to hit 30,000 miles per disengagement in 1 or 2 years. That slope will change unpredictably. I'd bet it gets steeper first, then slows down. But when, who knows?

That uncertainty of when is why TSLA is becoming a risky investment. If robotaxis are planned for production in 2 years, we are basically requiring progress to reach a high standard in a short window. Otherwise the produced cars are much less valuable.

Since Tesla doesn't provide disengagement data, the FSD Tracker critical disengagements is the best we have.

 

[Buckminster]

https://twitter.com/x/status/1776709972216443315

What is the Chinese interest rate?

 

[MD70]

I dont understand why there are concerns that there are insufficient batteries for a model 2? Its not like cheaper EVs than the model 3 have not been built by other companies. The MG4 exists, The Nissan Leaf Exists, there is a cheap peugeot EV. How are these other companies getting hold of batteries to make cheap compact EVs but Tesla cannot?
Also AFAIK Tesla are way ahead of any other car company when it comes to having control over their own battery supply chain, including giga Nevada, but also the 4680 line.
I know, I know... everyone is disappointed at the run-rate of 4680s right now, but we were all unhappy about FSD progress until recently. These things change fast!

As an investor, I'd like Tesla to announce and start making compact model 2s as soon as they can, EVEN if they are battery constrained to only make 100k a year. The fact that an affordable, small Tesla compact car exists will mean a lot of people make plans to get one as soon as the supply is available. Obviously from a mission POV, all EV sales are good, but as an investor I'd rather people put a reservation in for a M2 than an MG4 or Leaf .

Also, a small compact car is a new segment for Tesla. Not a bad idea to get a vehicle out there in customers hands, even in (relatively) small numbers so it can be perfected before it gets made in the millions?

I would not be surprised to see robotaxi AND model 2 announced and demonstrated on the 8th.

Battery’s are available if planned for, but tesla doesn’t seem to have planned for sufficient numbers in the next 2/3 years.

 

[GhostSkater]

Today was the last day with “my” Model 3 of three weeks that turned into a month

You guys won’t believe how many kWh I used, maybe we do live in a simulation after all

Amazing experience and hope to own one one day, or a Model Y with a custom portal knuckle and air suspension that I started to sketch out and do some napkin math, easily +4” to +6” ground clearance when needed while keeping the factory ride height and handling when on pavement, add some A/T tires and it would be the ultimate stealth off roader

I have one complain, please don’t exile me for it, I miss the lack of audible feedback. I’m not saying I miss the ICE sound, I don’t, even been building and riding small electric vehicles for the past 10 or so years as a hobby and professionally, and all of them have a bit of speed and power dependent noise. The three has, but it’s so quiet that you can barely hear in most situations

Cybertruck videos shows that Tesla did put an effort into that for it, which makes sense with the target demographics

The rest top notch, no comments (maybe the wireless charger not having any cooling and overheating my phone and barely charging it, but I think it’s more of a phone problem than the car)

Already rented for 2 more weeks a month from now for a longer trip

 

[Carl Raymond]

Maybe the Robotaxi is the van not the M2. You heard it here first.

Robo minibus will need to exist to be of benefit in zones where congestion is the limiting factor.

But don’t forget Musk responded with to Sawyer Merritt’s post about same platform for compact n robotaxi. If the one line makes both, switching at will, it protects Tesla from regulatory hurdles. They can keep winning with either product.

Been thinking about how FSD for people leads to FSD for goods, with humanoids assisting at the end points. Massive economic advantage to blocs who perfect this. Considering USA/China/Europe, none has a choice to not adopt. Economics drives it. Europe does not appear to have realised this yet.

 

[dhanson865]

Did you get FSD? Why not if no. How much did you save by no FSD.

Seems like used M3 with FSD are a good bargain now. 25K?

Maybe even MS with dead battery is worth 12K for the FSD transfer.

The car was $24,600 before tax credit and the tax credit has a cap at $25,000. I couldn't get FSD and still get the $4,000 instant credit.

The car was delivered with FSD 12.3.3 installed. I don't know if the prior owner paid for the FSD computer upgrade or if Tesla did it between owners.

Cameras just finished syncing enough to allow regular autopilot. Should be able to enable Supervised FSD later today or sometime tomorrow.

I got the 30 day FSD trial staring on the day of delivery even though the cameras weren't calibrated.

I did not get a FSD demo from the delivery center staff.

The car does still have the Intel Atom processor and has one scratch on the paint job, oh and no heated steering wheel. Other than that it seems pretty much perfect.

Oh and they did NOT give me a mobile connector or a USB drive. I had bought a mobile connector in a prior year to charge my Nissan leafs (using an adapter) so I'm not hurting there but I would like to have had a 2nd one. I ordered a USB drive from Amazon so I'll be able to enable shutdown sounds and sentry and such in couple of days.

 

[cliff harris]

But don’t forget Musk responded with to Sawyer Merritt’s post about same platform for compact n robotaxi. If the one line makes both, switching at will, it protects Tesla from regulatory hurdles. They can keep winning with either product.

I thought the most up to date info on this was isaccson's bio, which IIRC was clear that compact & robotaxi would share a production line, with talk that sterring wheels and pedals would be designed to be easily removed later.
Frankly I think thats a bit idealistic, but I can see Tesla having the exact same body shape and majority components for both. Maybe thats what we get on 8/8? A model 2 reveal with a 'robotaxi variant' being the exact same vehicle apart from controls.

 

[rentierparasit]

I dont understand why there are concerns that there are insufficient batteries for a model 2? Its not like cheaper EVs than the model 3 have not been built by other companies. The MG4 exists, The Nissan Leaf Exists, there is a cheap peugeot EV. How are these other companies getting hold of batteries to make cheap compact EVs but Tesla cannot?
Also AFAIK Tesla are way ahead of any other car company when it comes to having control over their own battery supply chain, including giga Nevada, but also the 4680 line.
I know, I know... everyone is disappointed at the run-rate of 4680s right now, but we were all unhappy about FSD progress until recently. These things change fast!

As an investor, I'd like Tesla to announce and start making compact model 2s as soon as they can, EVEN if they are battery constrained to only make 100k a year. The fact that an affordable, small Tesla compact car exists will mean a lot of people make plans to get one as soon as the supply is available. Obviously from a mission POV, all EV sales are good, but as an investor I'd rather people put a reservation in for a M2 than an MG4 or Leaf .

Also, a small compact car is a new segment for Tesla. Not a bad idea to get a vehicle out there in customers hands, even in (relatively) small numbers so it can be perfected before it gets made in the millions?

I would not be surprised to see robotaxi AND model 2 announced and demonstrated on the 8th.

Hi, Cliff --

>I dont understand why there are concerns that there are insufficient batteries for a model 2?

The world is currently awash in batteries; the concern is IRA eligible batteries. I don't follow this closely,, but Troy thinks that oncoming sources of supply are already spoken for. It's possible that self-help via 4860 could adress this.

Yours,
RP

 

[Musskiah]

A critical disengagement happens when the AV is about to hit something, and the driver intervenes.

This is false. A Critical Diengagent COULD occur when the *EV is on a collision course, but more often it is simply associated with an unsafe maneuver such as running a stop light or driving on the wrong side of the road or as @Discoducky reported still a common issue: driving past a "ROAD CLOSED" sign. This can happen even if no other human or vehicle is around, and it often does.

The fact that miles / critical disengagements went from 100 to 400 in the last 6 months doesn't mean it's going to hit 30,000 miles

Now that we have established Critical Disengagements (CD) DOES NOT EQUAL collisions, there is absolutely no reason 30k miles between CD needs to be achieved. In most instances, the car could pull over and/or a human could likely remote in and solve the issue. This could realistically occur mostly and wouldn't be mission critical.