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

[hobbes]

https://twitter.com/x/status/1621761294096375808

 

[SageBrush]

Current 4680 capacity is around 1/3rd of what would be needed simply to handle all the US MYLR cars. It's not even close, even without the CT eating them up.

Sounds about right, although I view the data very optimistically since the ramp has been going nicely the last 6 months; and I'm not even sure that Tesla-Austin has the capacity yet to build more Model Y than current 4680 production. By the time the CT line is ready to produce in volume, I think there is a good chance that 4680 production will be available.

 

[StealthP3D]

If you think Elon is envisioning a van stamped out of heavy, soft and rust-prone mild steel, like all other vans on the road, you haven't been paying attention to what's going on with the Cybertruck (CT) prototypes we have seen multiple examples of as it goes through its development process. The first two prototypes we saw had cold-rolled stainless-steel panels that were clearly bent on a traditional metal bending brake. I believe it was Elon who even commented on the need to score the metal along each bend before it was bent.

However, in the last several months every new prototype we have seen has had a distinct waviness to the body panels that can be clearly seen when the reflection in the panel is viewed at an oblique angle. Maybe someone who knows more about cold-rolled stainless steel than I do can comment but I know enough about the properties of cold-rolled stainless steel to know that it is not naturally wavy and bending it upon straight lines is not going to make it take on a wavy appearance, even if done imperfectly. And the initial prototypes, while they did have misaligned body panels, there was not a hint of waviness visible in their reflective surfaces. This is a super important clue that the production process has changed in a fundamental manner, a production process that I have not seen a single other person contemplate and one that has far-reaching implications for the economics of Cybertruck production and every Tesla utility vehicle hereafter. Maybe even every new model that Tesla releases from here on out.

I think the waviness of recent CT protypes tells us something big has changed in the way these prototypes have been manufactured. I suspect Tesla is buying cheaper rolls of soft, hot-rolled stainless steel and cold working them in-house into hardened CT panels. This would mean they are no longer using a metal brake but bending panels from soft metal on a long line of custom hydraulic rollers as they are also hardening the panels under high roller pressure. This would require development of many hardened rollers of varying angles, so the soft metal panels are progressively bent and hardened they travel down a long line of rollers. Most of the bending would happen quickly at the beginning of the line and most of the hardening and final bending to tight tolerances would happen after that. Ultra-hard, high precision panels would be ejected at high speed at the end of the line, ready for welding into the form of Cybertruck. Unlike an industrial metal brake this would be a continuous process.

Advantages:

-Panel details and cutouts could be performed on soft metal making it much cheaper and quicker to cut. If the metal came from the factory in full-hard form, every shaping and cutting operation is more difficult and requires more expensive, generally slower tools, non-laser/plasma tools wear quickly.

-The grain structure of the hardened metal can be controlled somewhat to optimize strength in the desired areas, in the desired directions. This would be done by adjusting the progression of the rollers to suit desired directional strength.

-The panel thickness can be controlled locally. Welding flanges can be left thicker and the panels can be rolled thinner as required to reduce weight and increase efficiency of material utilization, thereby reducing cost.

-Cost of material is reduced by buying the stainless steel closer to raw form, in its soft state and cost of manufacture is reduced by greatly simplifying production (once such a line is developed and running).

Disadvantages:

-Huge development costs and risk of delays. What I'm suggesting is not easy and has never been done before, AFAIK. It's essentially bringing the cold-rolling hardening process of a metal foundry in-house and integrating it with a roller bending/forming operation. While the potential benefits are mind-blowing, the sheer amount of trial and error needed to get this right shouldn't be under-estimated and I'm not sure it could even be done in a two-year period. So that would seem to argue against this analysis.

-Appearance. If Tesla cannot tune the waviness out of the reflection, it is not as appealing to the eye.

The only other thing I think could cause the waviness is if Tesla adopted a different welding technique that is warping the panels. But I'm not sure why the early prototypes did not have this appearance. Elon's recent comment about a people/cargo van makes me think something about the Cybertruck production/economic projections have been going exceptionally well and he wants to make more vehicles using CT technology!

I'm just dying to learn more about the unique challenges CT presents and the kinds of solutions being applied. It's driving me nuts actually. I don't envy anyone doing financial projections on Tesla because there is still so much unknown. We live in exciting times!

Last summer I posted the observation above, an observation that was precipitated by the sudden appearance of Cybertruck prototypes that had a waviness to the panels that was apparent when reflections in the panels were viewed obliquely. The initial prototypes, I believe including the one that was first displayed at the Petersen Automotive Museum, didn't have this characteristic visible. Also, that Elon was Tweeting about a highly configurable utility/passenger van at this time. I suspect he wants to migrate the Cybertruck chassis/body technology into new models.

Videos of the most recent Cybertruck pre-production examples still have this characteristic waviness to the flat panels but it looks more uniform and better controlled. I still think it's due to an entirely new production process being developed by Tesla that increases their vertical integration by using a cold-rolling process to harden the panels as they are bent and formed in order to reduce raw material costs and avoid the issues of having to fabricate the panels out of steel that has already been hardened by cold-rolling at the metal foundry.

The waviness could also be caused by stresses introduced in the panels as they are welded to one another, but I think this the less likely scenario since the earliest prototypes did not appear wavy.

I've quoted the post above for those who may have missed it last July. I'm super stoked about the Cybertruck going into production, not so much because of its clean, simple look, which I am also quite a fan of, but because of this revolutionary new production process that takes modern unibody vehicles to an entirely new level, resulting in a utilitarian vehicle, without paint, that is going to last and last (regardless of the process used to fabricate them). Over the next year or two, keep a close eye on how the economics of Cybertruck manufacturing evolve. Like all new technologies, it's probably going to take a certain amount of time and innovation to drive costs of the manufacturing process lower, the question is, how big is the potential prize at the end and will that usher in an automotive revolution of lighter, stronger and more durable automobiles that cost less to make in great numbers?

 

[SageBrush]

(graphic)

That was before the price cuts and subsequent partial small price increases of Q4/2022 -- Q1/2023

Seems like pre-history

 

[ZenRockGarden]

Sounds about right, although I view the data very optimistically since the ramp has been going nicely the last 6 months; and I'm not even sure that Tesla-Austin has the capacity yet to build more Model Y than current 4680 production. By the time the CT line is ready to produce in volume, I think there is a good chance that 4680 production will be available.

Oh, I expect most high volume Tesla models to eventually go 4680. For production cost reasons, not because it works better.

And I'm more optimistic about Tesla getting the CT rolling in volume than I am about sudden leaps in 4680 cell production. Both will come, but the CT ramp is the easier one imo, and so it's gonna need a big "reserved" sign on whatever supply they can make of 4680 till that ramp exceeds CT production, at which time there might be a model 2 looking for high-volume cheap-to-manufacture cells....

 

[StealthP3D]

That is way too simplistic. Day trading speculator bought on the tweet, and lost money when the deal collapsed and they sold. <<<shrug>>>

I personally also bought on the tweet. The only difference between me and the aggrieved speculators was that I held on to the stock.

I don't believe aggrieved speculators instigated the lawsuit, they were sniffed out by lawyers working for people who wanted to harm Tesla/Elon Musk. Of course, they are going to join the suit when told they don't have to pay for the attorneys if they lose.

 

[SageBrush]

And I'm more optimistic about Tesla getting the CT rolling in volume than I am about sudden leaps in 4680 cell production.

CT in volume and 4680 in volume have been linked from the beginning since the presumed 4680 scale and COG were to enable the CT price points.
Unless Tesla does an about-face and decides to sell an $80k CT, I think the earlier plans are still in place.

Which means that CT in volume implies 4680 in volume

 

[SageBrush]

I don't believe aggrieved speculators instigated the lawsuit, they were sniffed out by lawyers working for people who wanted to harm Tesla/Elon Musk.

That smells of conspiracy crapola. You don't need TSLAQ trolls to dig up lawyers willing to sue Tesla for $Billions on contingency. It could, however, make sense for a capitalized hedge fund that shorted TSLA to front enough money to keep a decent class action law firm in operation for the 5+ years needed to get to trial. I don't know of any evidence that happened, but it would be an alignment of shared interests.

 

[MC3OZ]

CT in volume and 4680 in volume have been linked since the beginning since the presumed 4680 scale and COG were to enable the CT price points.
Unless Tesla does an about-face and decides to sell an $80k CT, I think the earlier plans are still in place.

Which means that CT in volume implies 4680 in volume

They are probably working on things in parallel:-

• New improved cells and testing the cells.
• Line speed and yields.
• Ramping additional lines and training staff
• Cathode plant
• Raw materials.

They have said CT in low volumes in 2023, and 2024 should be volume 4680 production.

The 2023 CTs will probably be higher trim models, or they will pick the best options.
Having enough motors for 3 or 4 motor versions seems possible.

 

[FutureCTOwner]

I haven't seen this discussed here, but if it has apologies for doubling up.

What are people's thoughts about the possibility that Tesla may announce more than one Gigafactory on investor day? Mexico.... AND Indonesia? Both rumours seemed credible?

 

[MC3OZ]

I haven't seen this discussed here, but if it has apologies for doubling up.

What are people's thoughts about the possibility that Tesla may announce more than one Gigafactory on investor day? Mexico.... AND Indonesia? Both rumours seemed credible?

South Korea is another location that has been mentioned.

I think more than one location is possible.

It is also possible that Indonesia is only making cells, and that Mexico and / or Korea are only making cars. Korea could use some cells from Indonesia and some from Korean cell makers.

Mexico and Korea have car industry experience and parts suppliers.

 

[NickFie]

You cannot listen to an answer to a question that was never asked.

I beg to differ. Paradigm shift sensitivity has contributed to my career and “mad money” investment success. My current favorite is the way Munro & Associates - a deeply reality-based and pro-innovation organization - models the automotive OEM blind spots on Tesla’s Agile Manufacturing.

What I think OEMs don’t see - key practices that support light-speed innovation on a running assembly line:

1. Complete vehicle instrumentation with self-tune, self-test and report capability. Similar to completely automated software test and reporting in mature Agile DevOps.

2. Completely automated plant monitoring, analysis and reporting. This keeps everyone on the same page for status and priority.

3. Digital Twin capability. Tesla knows EXACTLY which parts and versions are in each vehicle. This streamlines long-term tracking and customer service.

4. Multi-cell parallel production steps. Changes can be implemented in a single cell and slip-streamed into production with less disruption. Similar to A/B website tests.

Please pardon my YouTube link omissions. Sandy Munro and Corey don’t see these and don’t ask.

- Munro & Associates interview Joe Justice. They don’t know to ask. Joe Justice may not recognize the blind spot.

- Fazad Mesbahi and Corey discuss Tesla cost advantage. Corey can’t explain why FSD hardware suite is in every vehicle. See #1 above.

- One of Sandy’s heads-up innovation videos. He attributes Tesla’s rapid cycles to deep pockets. Cannot see the deep automation and cells called out above.

- Even Connecting the Dots misses the cell structure in the Agile at Tesla series.

Apologies for lack of links. Time pressure.

 

[petit_bateau]

They are probably working on things in parallel:-

• New improved cells and testing the cells.
• Line speed and yields.
• Ramping additional lines and training staff
• Cathode plant
• Raw materials.

They have said CT in low volumes in 2023, and 2024 should be volume 4680 production.

The 2023 CTs will probably be higher trim models, or they will pick the best options.
Having enough motors for 3 or 4 motor versions seems possible.

We can do some interesting numbers with the info we have.

Let us take the 1.8m guideline as being the 2023 reality (which is a big if, I'll return to it).

Let us assume that Fremont and Shanghai both remain flat, neither up nor down. Big simplification.

Then since the Q4-2022 production rate was 1,758,804 on an annualised basis, which is near enough 1.8m and means they already have that cell capacity and vehicle assembly capacity in the system to deliver the 2023 guideline. (I am making the simplifying assumption that there was no cell stock build or drawdown).

On the call I cannot remember if they said it was 1 of 4, or 1 of 3, of the new 4680 lines running at Austin. So let us say that 1/3 of the Y production at Austin is from 4680 and 2/3 are currently 2170. This means that if the Austin production rate remains constant through 2023 and the other 2 x 4680 lines go live at the same throughput rate, then Austin will be able to completely eliminate 2170 usage, liberating those 2170 for the Semi.

My estimated apportionment of Q4 production to lines/sites was 26,596/qtr of the Y from Austin. At 75kWh per Y that is 1,994,700 kWh/qtr of cell usage. Using the 2/3 factor that is 1.3GWh/qtr of 2170 that would progressively become spare during the year, assuming no Austin Y production increase. If there was a unrealistic step change to all becoming spare on 1-Jan-23 then that is 5.3GWh of 2170 capacity in the full year. (Assume no increase in 2170 throughput either). Assume a triangular ramp over the 12-month period and that becomes a 2.6 GWh of freed up 2170 capacity during 2023 and 5.3GWh of 2170 capacity in 2024.

The Semi cell usage is somewhat murky, and I have seen both 900 kWh and 1,000 kWh stated. The demonstrated achieved range implies at least the 900 kWh and so I will assume 1,000 kWh so that there is some safety margin. Assume only Semi-LR are made in 2023 and 2024 for now. (I am holding the decimals in my spreadsheet, rounding in these words).

2023 : 2.6 GWh / 1,000 kWh / Semi-LR = maximum production of 2,660 Semis in 2023 using 2170
- interestingly I had 770 in my Semi ramp forecast for 2023; so perhaps an underestimate
2024 : 5.3 GWh / 1,000 kWh / Semi-LR = maximum production of 5,319 Semis in 2023 using 2170; at least until the new 4680 plant at Sparks comes onstream
- making the assumption that those 4680 then liberate further Panasonic-2170 usage from elsewhere (Fremont 3 and Y) for Semis
- interestingly I had 18,000 in my Semi ramp forecast for 2023; so likely an overestimate, though that does depend when the new facility comes onstream (! )

Let us assume that Cybertruck volumes are trivial in 2023, so 4680 consumption in 2023 by the CT is not material. Let us then consider the matter of the Austin and Berlin Y production. The above calculation assumed that Austin Y production remained flat at 26,596/qtr and simply transferred over from being 1/3x 4680 and 2/3 x 2170 to become 3/3 x 4680.

26,596/qtr x 4 = 106, 384 Y from Austin in 2023

similarly, assuming flat usage of China LFP and LG 2170 in Berlin:
36,000/qtr x 4 =144,000 Y from Berlin in 2023

However it is also clear from Musk's blurt on the call that as a minimum the real internal target is 2.0m cars/year in 2023. I noted that someone has started using the term cars in the call, not vehicles, so there is now an attempt to separate out the Semi numbers somewhat. We still haven't seen an explanation of why they don't appear in the P&D or Qtly reports, so that in itself is informative.

Nonetheless if Shanghai and Fremont both run steady in 2023 then that extra 200k must come from somewhere, most likely split fairly evenly between Austin and Berlin.You'll see where I am going in a minute So for 2.0m in 2023 we get something like:
- Austin Y = 106k + 100k = 206k in 2023, all of which will be on 4680 by year-end, and note this now includes increased throughput rates on all the Austin 4680 lines
- Berlin Y = 144k + 100k = 244k in 2023, all likely on a mix of LFP and LG-2170

Interestingly my own ramp forecast for both facilities (using previous ramps a guideline) were for 370k Berlin Y and 290k Austin Y, hence my own estimate for 2023 of 2,279, 820. If the 2.0m Musk blurt is more likely this suggests that the ramps are in fact cell-constrained through 2023.

Anyway what it also says is that :
- at end 2022 Austin 4680 is 26,596/qtr x 75 x 1/3 = 664,900 kWh/qtr = 2,659,600 kWh/yr = 2.6 GWh/yr
- at end 2023 Austin 4680 is 76,596/qtr* x 75 x 3/3 = 5,744,700 kWh/qtr = 22,978,800 kWh/yr = 23 GWh/yr

* note triangular ramp assumption again to get 100k extra out during the year, which would be a low ramp by Tesla historicals

And now for the really interesting bit. Therefore the Austin 4680 production should have ramped from 2.6 GWh/yr to 23 GWh/yr in 12-months of 2023, a 20 GWh of production add. And it will know exactly what to do to repeat the same trick again in 2024, assuming that the long lead machinery for the cell lines is in hand, and assuming that the raw materials are coming through the pipeline.

If we then look to 2024, and make the simplistic assumption that indeed another set of 4680 lines go in that are dedicated to the Cybertruck, (and that Austin Y production now remains flat, and that 460 ratesmachine are also by then flat), then in their first year they should make 10 GWh of extra 4680 cells. Assuming the launch versions of the Cybertruck are the rumoured 250 kWh version that means there will be cell capacity for 40,000 Cybertrucks in 2024. Up until now I was assuming a more aggressive ramp with 100,000 in 2024 after 7,000 in 2023. (Sometimes I am an optimist, looks like that 100k is being too much of an optimist). If 40k in 2024, and if they continue to run with 250kWh packs, and continue to build out line capacity, then that becomes 120k in 2025.

So, some interesting numbers.

 

[Artful Dodger]

I haven't seen this discussed here, but if it has apologies for doubling up.

What are people's thoughts about the possibility that Tesla may announce more than one Gigafactory on investor day? Mexico.... AND Indonesia? Both rumours seemed credible?

Very likely, IMO. The leading candidates include: (in alphabetical order)

1. Indonesia (nickel/battery plant)
2. Montreal, Canada (cathode materials?)
3. Nuevo Leon, Mexico (Parts manufacturing, Gen. Assb'ly)
4. South Korea (Possible Gen 3 platform site)

All four sites could be announced on Investor Day. Conspicuously missing? A new China site. I see S. Korea being chosen as an alternate to a 2nd China factory, due to a combination of health policy and geopolitical considerations. Of course, I have no special insight into all the issues at hand, but at China speed, it seems very likely a domestic Gen 3 plant could be created quickly sometime down the road if desired.

The above list does not include, but implies, continued expansion at Giga's Texas, Nevada, and Berlin (adding new assembly buildings, production lines, w. modular 25 GWh/yr battery lines #37 - the 'G-Cube'

Cheers!

 

[TN Mtn Man]

I imagine this has been covered, but I didn't see it.

Cadillac Lyriq Now Qualifies For EV Tax Credit Following Updated Classification Standards​

https://gmauthority.com/blog/2023/02/cadillac-lyriq-now-qualifies-for-ev-tax-credit-following-updated-classification-standards/

It didn't take the administration too long to rewrite aspects of the bill to take care of Mary B and the campaign donors in the UAW. Just long enough for Tesla to commit to price cuts. Now, everyday compact crossovers are "SUVs" even though they have no off-roading and minimal towing capability. The plus, as long as Tesla is still making a solid profit at their new pricing, their demand stays high. And of course, given GMs inability to build the Lyric in any volume, even people that might want one can't get one anyway.

 

[Prunesquallor]

…
I read that it took the jury two hours to reach a verdict. Two seconds should have been enough.

There’s some initial paperwork, then they have to choose a foreperson. After that they probably took one poll, saw it was unanimous, shrugged, and called in the bailiff. A little more paperwork, done. Two hours seems about minimum.

 

[OrthoSurg]

https://twitter.com/x/status/1621761294096375808

Does this include ICE cars for GM, Ford and Hyundai or this infographic is only about EV profitability?

 

[trayloader]

I imagine this has been covered, but I didn't see it.

Cadillac Lyriq Now Qualifies For EV Tax Credit Following Updated Classification Standards​

https://gmauthority.com/blog/2023/02/cadillac-lyriq-now-qualifies-for-ev-tax-credit-following-updated-classification-standards/

It didn't take the administration too long to rewrite aspects of the bill to take care of Mary B and the campaign donors in the UAW. Just long enough for Tesla to commit to price cuts. Now, everyday compact crossovers are "SUVs" even though they have no off-roading and minimal towing capability. The plus, as long as Tesla is still making a solid profit at their new pricing, their demand stays high. And of course, given GMs inability to build the Lyric in any volume, even people that might want one can't get one anyway.

The Lyric will be an ugly, costly Flop, looking like from the 60s.
The maximum the clusterf**£ GM is creatively capable of.

Let’s analyze the sales numbers here end of 2023, when Tesla announced a real SUV.

 

[Artful Dodger]

If you're going to look at previous history, then the odds are...

Fun Fact:​

I estimate that TMC's Perpetual Investors' Roundtable will reach* Comment #420,069 on Wednesday, April 26, 2023​

#Predict

*Bonus pts for the hex2dec.

 

[TesslaBull]

Even with color Blue, white interior, and 19" sport wheels the lease payment is only $386/month. That's less then the basic 3 just yesterday.

I still don't understand why anyone would literally rent a car for $500/month for three years and be left with nothing, and just have to give the car back at the end of the rental period???????