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

[Acroyear]

Been cleaning up my hard drive and came up with this gem. UBS Model 3 Teardown report

View attachment 959687

View attachment 959688

Lol, this reminds me of all the analyst reports that used to say Apple was going to end up last in smartphone market share, after Android and Windows Mobile(?!?). One of the reasons I invested in Tesla: )

 

[hobbes]

Do we know how far the die shrink went on the HW 4 FSD chips?

Has the Tesla AI gone so far that giving it coaching is more promising than training the NNs now ?!

 

[JupiterMan]

Been cleaning up my hard drive and came up with this gem. UBS Model 3 Teardown report

View attachment 959687

View attachment 959688

Goes to show that no matter how big the organisation, how shiny their offices and how long they have been trading, it is always just an opinion, and can be completely wrong.

That markets give such organisations high weightings for their models/valuations, and let these analysts dictate buying/selling decisions is frankly incredible.

But always remember, had these fools and the rest of the market not got Tesla so wrong (and continue to do so) then there would be less stock return on the table for the rest of us.

For that I am thankful.

 

[mongo]

1st clue that this was a hoax should have been when a UK-based CT reservation holder gets an early delivery. UK, Euope, Asia, OZ (and even likely Canada) will all have to wait. It's the IRA battery production tax credit, you know. I estimate the Gov't will toss around $45 * 200 = $9K at Tesla for each CT they sell in the USA. Now that's in addition to the Federal and State incentives that the purchaser gets in the USA. There might be 10 years worth of backlog for CT for this spec. Europe and Asia may have to wait for their own factories before they get this vehicle.

Cell and module manufacturing is based on US manufacturing, not US sales. Giga Mexico may not get local cell production for a while.

SEC. 45X. ADVANCED MANUFACTURING PRODUCTION CREDIT.
(a) IN GENERAL.—

(1) ALLOWANCE OF CREDIT.—For purposes of section 38, the advanced manufacturing production credit for any taxable year is an amount equal to the sum of the credit amounts determined under subsection (b) with respect to each eligible component which is—​

(A) produced by the taxpayer, and​

(B) during the taxable year, sold by such taxpayer to an unrelated person.​

...
d) SPECIAL RULES.—In this section—

(1) RELATED PERSONS.—Persons shall be treated as related to each other if such persons would be treated as a single employer under the regulations prescribed under section 52(b).​

(2) ONLY PRODUCTION IN THE UNITED STATES TAKEN INTO ACCOUNT.—Sales shall be taken into account under this section only with respect to eligible components the production of which is within—​

(A) the United States (within the meaning of section 638(1)), or​

(B) a possession of the United States (within the meaning of section 638(2)).​

(3) PASS-THRU IN THE CASE OF ESTATES AND TRUSTS.— Under regulations prescribed by the Secretary, rules similar to the rules of subsection (d) of section 52 shall apply.​

(4) SALE OF INTEGRATED COMPONENTS.—For purposes of this section, a person shall be treated as having sold an eligible component to an unrelated person if such component is integrated, incorporated, or assembled into another eligible component which is sold to an unrelated person.’’.​

If the Exoskeleton was abandoned, can we still hit it with a hammer? I would imagine the extra framing weight could be offset by thinner outer shell, but I hope not.

The shell is still a shell, a hard exoskeleton.
Perhaps a hermit crab.

 

[petit_bateau]

https://twitter.com/x/status/1683997914778030083

Tesla Energy FTW. Energy > Autos some day?

Over the past month, Tesla has secured contracts worth a combined $1.3 billion for its Megapack stationary storage battery product. And this figure only includes the ones we know/I've tweeted about. There are likely more that haven't been publicly disclosed yet.

Source

https://twitter.com/x/status/1683997914778030083

Tesla Energy revenues are approx $1.5bn/qtr for Q2-2023at a deployment rate of 3.6 GWh/qtr. For simplicity assume all this is Megapack out of Lathrop.

By Q4-2024 Lathrop needs to be at 10 GWh/qtr (i.e. 40 GWh/yr which is equivalent to 500,000 cars/yr) and revenues of $4.8bn/qtr, which is 18-months away and may represent steady state for a fully built-out phase 1 at Lathrop. At least that is my modelling.

Since Megapack lead times appear to be approx 24-months away then Megapack should as a minimum be booking orders in now at a rate of $4.8b/3 = $1.3bn per month. This is exactly what the tweet says is happening and it matches my ramp modelling.

So what that tweet is telling us is that phase 1 Lathrop is now fully booked at max rate production. So the full ramp should be done with in approx 18-24 months from now, and Tesla has confidence in that ramp rate if it is accepting orders at that rate.

I'm actually slightly more optimistic than that, and think they might achieve the full ramp in 12-months from now.

The interesting part is trying to figure out when/how Tesla can lay their hands on sufficient excess LFP to press the go button on the next Megapack facility. Whether that would be a Lathrop phase 2, or a Europe or Asia facility is another interesting question. And would they be able to do two simultaneously ?

IMO Megapacks produced at Lathrop will increasingly pivot to export due to IRA rules.

I think you are right that US LFP manufacturing may have an excess, but I'm not sure Tesla can safely base their North American Megapack production on that assumption.

Many of these factories are joint ventures with car companies, or have car companies as intended customers. IMO there are increasing signs that we are about to enter the "wild stampede" phase of EV adoption, where all car companies decide that swimming the moat ASAP is their only chance of survival.

But in particular I not convinced that any other form of IRA compliant LFP battery production can ramp faster that 4680 LFP, or produce cells on a cheaper per kWh basis than 4680.

Most of the the cost savings in battery day are "chemistry agnostic" the only exception is the 12% cost reduction due to Cathode material.

So I think it should be possible for Tesla to make 4680 LFP cells and that will probably be the cheapest LFP cells they can source on a per kWh basis.

Then we come to the question of whether or not 4680 LFP is adequate. For energy storage it doesn't matter, for vehicles the advantages of a structural pack probably cancel out any other minor disadvantages.

IMO the Austin cathode plant is important for the following reasons:-

• Allow 4680 production to ramp to high volumes
• 12% cost reduction per cell.
• Fully IRA compliant
• Single crystal cathode? (possibly later)

The Tesla long range semi is the vehicle that would benefit the most from using cathode material produced at Austin, especially if single crystal cathode is ever produced.

When the Semi needs to move to 4680 cells might depend in part of whether or not Sparks 2170 production can use cathode material produced at Austin.

An easy path to very high levels of IRA compliance is for Fremont, Sparks and Mexico to source battery raw materials from Texas.

For the Berlin ramp, raw materials that are being imported into Sparks could be shipped to Berlin instead. I'm not aware of any current plan for a Cathode plant at or near Berlin.

Class A and Class B cells we don't really know what is going on with them, or why the grading is done. My suspicion is that Class A and Class B cells can't be mixed in a single pack, but an all Class B pack is fine. Tesla just needs to avoid shipping a vehicle with a Class B pack to a customer?

Interesting points, thank you.

I do not understand if there is any added value in putting LFP in a 4680 can vs putting it in a prismatic format.

And isn't prismatic LFP sufficiently structural as to be effectively interchangeable with a 4680 pack in mechanical terms (though really that question is only really an issue for the automotive applications).

If 4680 has no particular advantage for LFP, then why would Tesla split its 4680 effort across two chemistries ? It seems to me that Tesla's 4680 effort is going to be fully occupied in trying to ramp 4680 as the high-end automotive solution. Leaving LFP as the low-end auto solution and the stationary solution (until such time as sodium/etc arrive one day maybe).

Can you or anyone else answer those question ? For example @GhostSkater or @M3PMike any of the other battery afficionados ?

 

[jerry33]

I do not understand if there is any added value in putting LFP in a 4680 can vs putting it in a prismatic format.

Prismatic cells have an inherent problem vs cylindrical cells. The radius of prismatic cells varies depending upon where you take the measurement. This leads to uneven heating. Almost all Prius cell failures were due to this. I don't believe the cell chemistry alters this, although some chemistries may be less problematical by a small degree than others.

 

[Usain]

You are correct she said Ultium technology. This is all about cost. I agree there is nothing special about Ultium but it is also a big insourcing project as they realized if they don't do this they have no chance to be competitive.

They are likely completely upside down on cost with the current Bolt by buying everything from LG. Why else would you stop production of your highest volume EV? They are shipping each one with $$$ in the trunk as it likely costs more to make than the current selling price.

The only way they are going to have a chance to be cost competitive is stop buying from LG and use all their internal components, cells, pack, BMS, drivetrain etc. Ok, the cells are still a joint venture with LG but Ultium cells are also made in the US so they also will get the full $45 Kwh IRA credits for cell/battery manufacturing which the current Bolt does not get with manufacturing in South Korea. This is separate from the $7500 consumer credit which has different sets of rules. (Car made in NA, battery and material sourced from US or free trade country.)

We broadly agree.

I do want to correct the record with regard to South Korea. The current Bolt (non-Ultium) battery packs are now manufactured here in the US. This is the pack I expect GM to continue using in the Bolt for at least a few more years. It is GM's only reliable source of battery packs right now.

And that's why the Bolt is staying alive a little longer. Mary will add some "Ultium technology" to the old packs and call it progress.

 

[petit_bateau]

Prismatic cells have an inherent problem vs cylindrical cells. The radius of prismatic cells varies depending upon where you take the measurement. This leads to uneven heating. Almost all Prius cell failures were due to this. I don't believe the cell chemistry alters this, although some chemistries may be less problematical by a small degree than others.

Yes, I appreciate that. This appears to be one of the reasons why the various NMC/etc chemistries struggle with pouch formats, and (correct me if I am wrong in my simplistic view) ultimately a prismatic is just a pouch in a rigid box.

But for LFP there doesn't appear to be a shrink/swell cycle (excuse my simple terminology) associated with charge/discharge.

But for the lower-end automotive applications (approx 2/3 of vehicles) provided that the pack is sufficiently large (so no real thermal problem) then LFP prismatics appear to be satisfactory in all respects. And for stationary they appear to be satisfactory.

Hence my question as to whether there is any real benefit in putting LFP in a 4680 can vs putting it in a prismatic box.

 

[Cobbler]

Those AI generated articles are funny

 

[hobbes]

While it is possible to produce structural packs with other LFP form factors, we don't know if those packs would be a drop in replacement for 4680s.

IIRC Giga Berlin has already been using BYD blade battery packs for a while..

Was trying to find a source - pls link if you have one or correct me if I am wrong

 

[Buckminster]

Elon has just flown out to Italy. I am 80% confident he will be recording an All in Podcast in Portofino. it will air on Friday most likely.

 

[cliff harris]

An article headline I saw today:
Will Nikola Be Worth More Than Tesla by 2030?
I pity the poor misguided retail investors who do not do proper research. Also proves there are still a lot of future tesla investors who are currently being tricked by this nonsense. Eventually there will be enough Teslas on the street that people will stop believing such dross.

edit: gah! didnt refresh page, beaten to it!

 

[mongo]

I do not understand if there is any added value in putting LFP in a 4680 can vs putting it in a prismatic format.

There is if your manufacturing processes are set up around 4680s. Otherwise, you need to invest in prismatic equipment which would only supply one sub-group of products.
4680 have better thermal control, Megapacks run at C/2 for hours.

And isn't prismatic LFP sufficiently structural as to be effectively interchangeable with a 4680 pack in mechanical terms (though really that question is only really an issue for the automotive applications).

Prismatic are not made for axial load nor torque/shear. Twist a box vs twist a can.

If 4680 has no particular advantage for LFP, then why would Tesla split its 4680 effort across two chemistries ? It seems to me that Tesla's 4680 effort is going to be fully occupied in trying to ramp 4680 as the high-end automotive solution. Leaving LFP as the low-end auto solution and the stationary solution (until such time as sodium/etc arrive one day maybe).

Material availability, cost, fit for purpose.

Tesla earnings calls:

Q3:
Yeah. And on the cathode side, the [Inaudible] we think will probably be iron and most of the iron -- iron can scale to very, very high tonnage and then some nickel. The exact percentages are hard to figure out, but it's probably be twice as much iron cathodes as they [Inaudible]. And then there's the manganese wildcard as well [Inaudible]

Q4:
Yes. It's roughly -- cell supply is roughly matched with that. And the 1.8 million cars, if we get lucky, it could be more. And then, the rest would go into stationary storage, the Powerwall and Megapack.

 

[Atomsla]

Those AI generated articles are funny
View attachment 959708

This should've been a one word article.

 

[Chenkers]

The interesting part is trying to figure out when/how Tesla can lay their hands on sufficient excess LFP to press the go button on the next Megapack facility. Whether that would be a Lathrop phase 2, or a Europe or Asia facility is another interesting question. And would they be able to do two simultaneously ?

We already know where the next Megapack factory will be, because Tesla announced in April that they will be building it in Shanghai. Tesla To Build Megapack Factory In Shanghai - CleanTechnica This is for 40GWh/10,000 units per year with construction originaly expected to start in Q3 2023 and production in late 2024. However we don't know the exact location for this yet.

I expect that the batteries for this will come from the new 80GWh battery facility that CATL is currently constructing near Giga Shanghai. Wu Wa has been doing drone flyovers of this construction for months but it still appears a while away with construction ongoing on the worker accommodation apartments and factory buildings. This new factory is in addition to the current Shanghai plant that CATL operates from existing buildings at another location that currently supplies standard range Model 3 and Y from Shanghai. The new 80GWh plant was presumably sized for a compact China model but since thats been delayed around 50% of the total production will likely go into Megapacks. A recent Wu Wa video of the CATL construction is below.

 

[petit_bateau]

We already know where the next Megapack factory will be, because Tesla announced in April that they will be building it in Shanghai. Tesla To Build Megapack Factory In Shanghai - CleanTechnica This is for 40GWh/10,000 units per year with construction originaly expected to start in Q3 2023 and production in late 2024. However we don't know the exact location for this yet.

I expect that the batteries for this will come from the new 80GWh battery facility that CATL is currently constructing near Giga Shanghai. Wu Wa has been doing drone flyovers of this construction for months but it still appears a while away with construction ongoing on the worker accommodation apartments and factory buildings. This new factory is in addition to the current Shanghai plant that CATL operates from existing buildings at another location that currently supplies standard range Model 3 and Y from Shanghai. The new 80GWh plant was presumably sized for a compact China model but since thats been delayed around 50% of the total production will likely go into Megapacks. A recent Wu Wa video of the CATL construction is below.

Thank you, I missed that Tesla non-announcement. It is bonkers that Tesla did not even mention this in the Q2-2023 slide deck. Both of the links are real useful and the numbers look to be a cut'n'paste of Lathrop. I wonder if there will be a more formal announcement at some point, or if they are doing this without a fuss due to the China/USA political sensitivities.

I agree that LFP in prismatic form seems to be a drop-in replacement with sufficient structure vs either 2170 packs or 4680 packs. It is just a matter of making the box(es) mechanically adequate. So in principle the LFP supply could equally well into the 2/Z out of the Shanghai-annex. It certainly seems that the BYD LFP blades are going into the Y in Berlin.

 

[y_naught]

Been cleaning up my hard drive and came up with this gem. UBS Model 3 Teardown report

View attachment 959687

View attachment 959688

They nailed the uncertainty!

 

[scaesare]

Something about a robot that was trying to fluff a soft layer of insulation that was hard to grip...

Either that, or GM misunderstood what the purpose of a "Flufferbot" was....

Spoiler: Kinda NSFW

Urban Dictionary: fluffer

(noun) a person in the adult entertainment industry whose job it is to give male porno stars blowjobs in order to get them ready to perform.

www.urbandictionary.com

 

[Knightshade]

Tesla has settled another lawsuit (albeit for a lot less $ than the BoD one!)- this one over them jacking up pricing on solar roof for people who already had signed contracts at a lower price-

Tesla settles class action over solar roof pricing debacle

Tesla has officially settled a class-action lawsuit over a steep price hike on its solar roof that even affected signed...

electrek.co

Unrelated but interesting- the EU announced new fast charger requirements to greatly expand availability- including requiring chargers every 37 miles along all the EUs main transport corridors and requiring at-the-charger payment (which might be another puzzle piece for those contactless V4 pics we saw) without need for a subscription among other things detailed here:

EU passes law requiring EV fast chargers along all main transport corridors by 2025

The Council of the European Union (EU) has adopted a new law surrounding EV charging infrastructure it’s calling a “milestone”...

electrek.co

 

[Cobbler]

Either that, or GM misunderstood what the purpose of a "Flufferbot" was....

Spoiler: Kinda NSFW

Urban Dictionary: fluffer

(noun) a person in the adult entertainment industry whose job it is to give male porno stars blowjobs in order to get them ready to perform.

www.urbandictionary.com

Optimus could be a good flufferbot! Saw a video how it could touch an egg without cracking it