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

[Right_Said_Fred]

Just got the information an hour ago from a well known automaker that does not get enough battery cells/packs for the production hence they had to go down in output and the line is even partly standing while the supplier ask for a lot more money to restart delivery. Crisis meetings are ongoing here at Friday night, CEO is pissed as they did not tell him until a few days ago. Mid Management was shaking because they did not know how to convey the message.

Its what me and many other have predicted and it became true now. BTW, thats not public info and appreciate if that does not leave this forum.

Implications are extremely severe as I do not see that situation to change soon unless you have your own battery tech and production.

All I can hope for is that they start in large scale production in Europe and understand now how critical it is leaving their former wrong strategy. Its silly that they did not see it coming, incredibly dumb.

Tesla will in my opinion not benefit from that as their demand is going through the roof anyway and with lower output from other brands we have lower adoption rate from the broader consumer group.

With all these traditional OEMs struggling to get enough batteries, there is always this nagging question in my head: how did Tesla, in such a crowded and competitive market, get their hands on production capacity at LG and CATL for at least 150,000 Chinese Model 3s per year (and more when the Model Y plant comes online)? How are they snatching those batteries away in front of VW, Mercedes, KIA/Hyundai, who are all extremely battery constraint, and not pay their weight in gold? Does anybody have an explanation?

 

[Navin]

Have Index funds you cannot trade stocks, and you are waiting for TSLA S&P inclusion - look for funds that have TSLA.
Found one Vanguard Fund (VWUSX) that has TSLA as top 10 holdings. Surely they are others too.
cheers!!

My poor reading/dyslexia thought you wrote: (VWSUX). Haha

 

[Thekiwi]

Third Row Podcast Part 2 summary:

• Big change going on in the background of Autopilot: instead of having planning, image perception, etc all being separate, they're all being combined, with all camera imagery being handled jointly rather than running image recognition, etc on each one. Big change in both quality and efficiency.
• This is not Dojo; Dojo is about training speed. Comes out late this year or early next year.
• Sees computer vision as being like navigation was in the 1990s - a very hard problem due to the limitations of your hardware combined with early-generation software. Expects autonomy a decade from now to feel easy like navigation is today.
• Auto industry is used to slow rates of improvements. Still not really a car today that matches the 2012 Model S, from a price-for-capabilities standpoint.
• The founding principles of Tesla were basically completely wrong. Notion was we'd take a Lotus Elise, put an AC Propulsion drivetrain in it, but them together, we'll have an electric car, and it'll be great! Sounds pretty easy. But the AC propulsion technology basically couldn't be industrialized. Handcrafted electronics with an analog motor controller. It'd respond either completely or not at all. It just couldn't be scaled; it would have been basically finicky super-expensive prototype cars.Ended up using basically nothing of it. And basically 7-10% of the Lotus Elise. Would have been better just to make a car from scratch.
• But things are often like that. What matters is that you can iterate quickly and adapt.
• Mass manufacturing difficulty is underappreciated. Only at peak made 10 or so Roadsters per week. If two got made in day, that was a good day
• Final assembly was at an old Ford dealership in Menlo Park. The "test track" was just normal driving around the area.
• NUMMI plant was just a big box. Only broken equiment left. Some literally weren't even the scrap value. You couldn't get a scrap dealer to come take them for free. Tesla made a lot of them work, though (such as the plastic injection moulding machine, paint robots, etc). Body line had to be made from new, it was fully stripped.
• Worked out, but it was very difficult, esp. since top suppliers wouldn't work with them, or they'd get their "D-Team". It was vertically integrate or die.
• Tried outsourcing - battery was going to be made at a place that made BBQ grilles in Thailand. They realized this was a crazy idea, for all of the transport, how long it'd take to find out if there was a problem and iterate on it, etc etc etc. They could have ended up with 5 months of scrap inventory for a problem.
• There's no logic to having battery packs be made of modules anymore - they plan to move away from it. The initial idea, from the Roadster days was that if something broke, you'd swap out a module, like swapping out a dead server in a server room. But it's becoming less economically reasonable of a proposition now.
• Went through a lot of trouble designing the S pack for pack swap. But now even phones are going away from battery swap. Makes more sense just to increase the range and makes charging faster. It seems obvious now, but back then it wasn't as clear. But you still have atavism - just like Model 3 still has modules, Model S/X still have swappable batteries.
• The structure of an organization is reflected in the structure of the production. Tesla had a module team, so Model 3 has modules (this has been restructured to just a "battery team"). And then you end up with a box in a box. Because this team wanted an enclosure, and this team wanted an enclosure, so you have a box within a box. And then on the Model 3 you have the pack, and then an underbody, a third enclosure, because the body team wants to have an enclosed body. Lots of brackets on brackets. The teams historically haven't been integrated as much as he'd like to see.
• Building a production system is 100 times harder than a prototype. The reason that companies like Tucker died is because production is way harder than hand-building to show off.
• "Here's a real important point that is not well appreciated, this is a point that should be advanced by short-sellers, but I've not seen it articulated - but it should be: the incumbent car companies make most of their money by selling spare parts to their existing fleet at high margins. And they sell the new cars at a de-facto zero margin or even at a loss. It's kind of like printer cartridges and razor blades: you sell the razor at a loss and the blades at a profit, or the printer at a loss and the cartridges at a profit, or video game consoles - the actual cost of say, an X-box may be $600 but you can buy it for $300-400 because they make up for it on the games that are bought. So if you're a new company, you do not have a fleet. You have no fleet from which to subsidize the sale of your new cars. This is the primary reason there has not been a successful car company startup in the United States. Most car companies have 80%, 70% of their fleet out of warranty. ... Even if they stopped selling new cars, they would still... (laughs) their profits would increase!"
• Innovation rarely comes from insiders, who are complacent with their status; it usually takes outsiders. Such as with SpaceX: it's actually surprising how little innovation there's been in rocketry, despite SpaceX showing landing and reusing rockets many times.
• Not very warm on Rocketlab's idea of catching boosters with a helicopter, both from logistic and economic perspectives. But still thinks it's good they're working toward reusability.
• Earth-to-Earth: Main constraint is noise. Coming in for landing, the double sonic boom is very loud. Have to do offshore. But definitely can be done. And thinks that the economics can be made to work as well, to make it economically competitive with air travel.
• Electric VTOL aircraft: "a lot of difficulty associated with that". You can't do them all at once, with respect to allocating the resources. It can definitely be done, but it's where's best to put your resources, which are very constrained. Not just technically, but also, it's an entirely different regulatory regime. There are no car companies that are also aircraft companies.
• Asked if, due to the good financial health at Tesla, whether that means they can now start pumping a lot more focus on diverse R&D projects. Said it doesn't work that way. That if they know a factory that mass produces superb engineers, please let him know. It's not a money thing; there's a fundamental limit on "exceptional engineers" available.
• Asked about how priorities are chosen. They're usually dictated by necessity or choice. What do they absolutely need to make work E.g. with Model 3, there were so many mistakes with the production process that they had to dedicate pretty much the entire team to fixing Model 3 production problems. Said that it felt like the scene in Indiana Jones where they were being chased by the giant boulder
• Biggest focus now is to get to be making cars on each continent. Its insane to be making cars in CA and shipping them to Europe and Asia. Huge expense, high capital carrying costs due to at-sea inventory, tariffs, extra chance for damage, etc. Creates a lot of costs and is hard to manage. Increases factory complexity at California to deal with all different regulatory regimes. Including LHD vs. RHD, due to some random bureaucrat a hundred years ago And can't even have the same stickers, because not everyone speaks English.
• Having more factories will really destress the company a lot and improve its economics.
• Had to ship cars to Europe where the supplier of the headlights for Europe couldn't match rate, so they'd ship cars with US headlights and then swap them out at the port, just to keep the inventory moving.
• The first quarter last year was a tragedy of errors. Belgium (where they were shipping the cars) went on strike! "What do to you mean, Belgium went on strike?! Okay, now what do we do???" Also mentioned the China sticker issue.
• Ships are also a pain. Time waiting to ship, to load it... ships break down, get delayed by storms, get stuck waiting for ports, etc.
• Tesla has had a massive uphill battle to sell cars in China, due to no subsidies (unlike local manufacturers), facing tariffs, shipping, etc. Proud of how well they did despite the barriers they were up against.
• Cars in the US, they'd get paid by customers before having to pay suppliers, which is great for scaleup. For EU and China, it's reversed. Europe is much harder, due to the Panama canal delays. Esp. when the canal gets backed up. In the worst case they can be forced to go around Tierra del Fuego, which is a transit nightmare scenario.
• Asked what's the best reason for choosing Berlin rather than other places in Europe: (joking) "They have the best nightclubs" (everyone laughs). Says that they they investigated numerous locations, but needed to move quickly, and Berlin allowed that. The fact that BMW had been looking to build a plant there meant that there'd already been a year's worth of environmental paperwork been done on that location for an auto plant. Made it much faster. Local and state governments were also very supportive. Close enough to Berlin that people can still live in the city and commute to the factory. Train station will be moved to where you can hop right off the train and into the factory; no need to even drive in. Goes back to joking: "GigaBerlin sounds like the name of a great night club. I'd for sure work for a company that had a great nightclub"
• Asked about the roller coaster. Still wants to do that eventually. Would just have it be modified Teslas designed to ride rails.
• Asked whether he expected that many orders for Cybertruck. "No, no really?"
• Kimbal: "Elon told me that this was a daring design. I'm more excited about this design than any other design. And I saw it and I was just taken aback. Not by the design, just by the pure aggression that the truck.. you just stand in front of it, and it's like, "ooookay, I'm afraid". Elon: "Well, it seems that a lot of the reason that people buy trucks in the US, it's like, what's the most bad*ss truck. Like, which one's the toughest truck. Well, what's tougher than a truck? A tank. Like, a tank from the future!" ... "How do you out-tough a truck? You make a futuristic armoured-personnel carrier."
• "I wasn't sure whether nobody would buy it, or lots of people would but it. I told the team that if nobody buys it, we can make a truck that looks like every other truck".

Excellent summary.

I only wish there were some knowledgeable TMC members present - there were so many obvious followup questions to ask after some of Elon’s comments.

Elon focused a lot on how stupid it is to build cars in USA and shipping them to Europe and China, so one obvious followup is will the premium models eventually also see localized production in China and Europe?

(One note: Honda is one car maker that actually does make aircraft, The Honda Jet, but it’s tiny compared to its auto operations. I think Musk may start an independent company to make planes, by the way he was talking, but not any time soon.)

 

[dl003]

With all these big OEMs struggling to get enough batteries, there is always this nagging question in my head: how did Tesla, in such a crowded and competitive market, get their hands on production capacity at LG and CATL for at least 150,000 Chinese Model 3s per year (and more when the Model Y plant comes online)? How are they snatching those batteries away in front of VW, Mercedes, KIA/Hyundai, who are all extremely battery constraint, and not pay their weight in gold? Does anybody have an explanation?

I think if I was a supplier, I'd prefer to forge a relationship with an OEM who is going to be here in 10 years and learn from them. Battery making is not a one way street. There's a feedback loop between the OEM and the supplier.
Excellence attracts. You wanna play for the best team around.

 

[Skryll]

With all these big OEMs struggling to get enough batteries, there is always this nagging question in my head: how did Tesla, in such a crowded and competitive market, get their hands on production capacity at LG and CATL for at least 150,000 Chinese Model 3s per year (and more when the Model Y plant comes online)? How are they snatching those batteries away in front of VW, Mercedes, KIA/Hyundai, who are all extremely battery constraint, and not pay their weight in gold? Does anybody have an explanation?

Is that confirmed to be where they source their model 3 battery cells from?

One potential explanation - technology sharing. VW/Merc/Kia/Hyundai etc probably have nothing to share in that regards.

 

[mongo]

With all these big OEMs struggling to get enough batteries, there is always this nagging question in my head: how did Tesla, in such a crowded and competitive market, get their hand on production capacity at LG and CATL for at least 150,000 Chinese Model 3s per year (and more when the Model Y plant comes online)? How are they snatching those batteries away in front of VW, Mercedes, KIA/Hyundai, who are all extremely battery constraint, and not pay top prize? Does anybody have an explanation?

Tesla can use cylindrical (commodity) form factor cells. Other use pouch or prismatic. Cell providers are not investing their money into making OEM-centric form factors.
And it's not simple, most of that Model 3 ramp trouble at GF1 was related to the machinery to turn those cylinders into packs.

 

[BenPrice]

With all these big OEMs struggling to get enough batteries, there is always this nagging question in my head: how did Tesla, in such a crowded and competitive market, get their hands on production capacity at LG and CATL for at least 150,000 Chinese Model 3s per year (and more when the Model Y plant comes online)? How are they snatching those batteries away in front of VW, Mercedes, KIA/Hyundai, who are all extremely battery constraint, and not pay their weight in gold? Does anybody have an explanation?

I would guess volume.

Tesla will buy and use, with growth, many more than the rest combined almost.

 

[Krugerrand]

Its what me and many other have predicted and it became true now. BTW, thats not public info and appreciate if that does not leave this forum.

Um...

 

[Nuclear Fusion]

More fuel for Tesla? Stock could join the S&P 500 by the end of the year

At current valuations of around $100 billion, Tesla would debut in the top 70 names of the index. With a total market capitalization of about $28.1 trillion at the end of 2019 we could expect Tesla to command a weighting of about 0.35%.

That may not sound like a huge amount to a casual observer. But we all know a few basis points adds up in a hurry when you put enough zeros behind the numerator it’s applied to. Considering S&P’s own literature claims $9.9 trillion is benchmarked to the S&P 500, Tesla shareholders should salivate at the prospect of exposure to 0.35% of that figure that adds up to about $35 billion.

 

[AudubonB]

It is!
View attachment 506469

That's just silly.

No real company's stock price chart has ever or will ever look like that.


on edit: looks as though some Swedes don't get some Arctic humor........

 

[MP3Mike]

Just got the information an hour ago from a well known automaker that does not get enough battery cells/packs for the production hence they had to go down in output and the line is even partly standing while the supplier ask for a lot more money to restart delivery.

So the problem isn't that the supplier doesn't have a supply of cells it is that they want more for them? Or are they giving you the option of paying more to "steal" cell supply from another manufacturer?

 

[theschnell]

Just got the information an hour ago from a well known automaker that does not get enough battery cells/packs for the production hence they had to go down in output and the line is even partly standing while the supplier ask for a lot more money to restart delivery. Crisis meetings are ongoing here at Friday night, CEO is pissed as they did not tell him until a few days ago. Mid Management was shaking because they did not know how to convey the message.

Its what me and many other have predicted and it became true now. BTW, thats not public info and appreciate if that does not leave this forum.

Implications are extremely severe as I do not see that situation to change soon unless you have your own battery tech and production.

All I can hope for is that they start in large scale production in Europe and understand now how critical it is leaving their former wrong strategy. Its silly that they did not see it coming, incredibly dumb.

Tesla will in my opinion not benefit from that as their demand is going through the roof anyway and with lower output from other brands we have lower adoption rate from the broader consumer group.

If OEM had contract to buy a certain number of cells at a certain price, how can supplier change deal after the fact?

 

[MP3Mike]

If OEM had contract to buy a certain number of cells at a certain price, how can supplier change deal after the fact?

Probably at the risk of being sued. But really is suing them and getting your cells in a year or two going to help or do you just pay the extortion so you can get your factory running again?

Or maybe the contract had is opened ended enough that they can still meet it. (Delivery x number of cells in 5 years with the expectation of a consistent supply, and they want to sell current cells to someone else and then dump a massive amount of cells on them toward the end of the contract?)

IDK I'm just making stuff up.

 

[dakh]

By "crazy optimizations" did you mean "too high of an error rate"?

No not that, I'm not talking about the results of inferencing but about the process itself. Even with the cool new hardware there's still a limit on the size of the network(s) they can load and on the speed of inference. Musk alluded to that during the interview. With a few orders of magnitude leeway in compute 10 years down the road this might be easy, but with current hardware they might have to spend a bunch of time figuring out how to optimize for compute efficiency simply to make the system responsive enough or to fit all the brains they need into FSD computer. This isn't easy or deterministic from what I understand.

To put it in simple terms for non-technical folks, imagine trying to run current software on a computer that's 15 years old. It either simply won't fit, or will run so slowly that you're not gonna wanna use it. I don't know how close to limits of the new FSD computer are they going to get with what they need to run, I'm only saying until we know for sure we need to price that possibility in.

IMO the correct way to value Tesla's FSD program is as driver assistance where it has a real present value as opposed to an estimated value at some unknowable point in the future.

Looked at this way, a Tesla vehicle derives significantly greater value from its driver assistance than other vehicles and the lead appears to be increasing. This is a competitive advantage that seems likely to remain for the foreseeable future. No one else is trying seriously: Waymo and Cruise are about FSD, no driver assistance component. IIRC VW is trying to do something, but they still think software problems can be solved by hiring more developers.

That's a great point. I'm not sure if that's what Musk was talking about during the interview where he said you have to both be able to charge premium for the car, and to offer autonomy to successfully enter auto market. He didn't say "advanced cruise control", it was specifically "autonomy". Yet Tesla is doing just fine without full autonomy at the moment. I'm sure charging now for a promise of autonomy in the future is helping the bottom line or at least helping pay for AP development.

 

[avoigt]

So the problem isn't that the supplier doesn't have a supply of cells it is that they want more for them? Or are they giving you the option of paying more to "steal" cell supply from another manufacturer?

What I know is that they ask for a better deal. That indicates they either have supply or steal it from other customers. But the issues from purchasing going further back in time.

Its boiling up to the board now.

 

[dakh]

Tesla have been collecting full video and sensor data from the fleet for years for specific driving situations, i’m not sure if they also collected driver actions (steering, acceleration etc) but I’d be surprised if they didn't.

From what I remember, they were collecting image samples that either were manually or auto-labeled and they ran some stuff in the fleet that would look for specific kinds of images to send back to the mothership. I don't remember them ever talking about videos, do you have any links?

On the screenshots that are on the new AP page I can see they have pretty cool looking simulation, that is one way to collect a bunch of "videos".

Either way, I think we agree that some level of rework or new work is definitely there to be able to use the new system to good effect. That's all I really wanted to say, yes definitely this is progress, no I don't see this as being "almost done", far from that.

 

[Krugerrand]

I also want to take a moment to recognize this post by you on 12/20, sp=$400.

View attachment 506462

I did not believe it and did not take advantage. Well, my margin is at a level that is somewhat uncomfortable, so I have to fight my urges to buy more.

The post was barely more than a month ago.
What a ride since then.

1 like, 28 finnies, shame on us.

Don’t feel bad. How can you take anyone serious who’s so pompous as to say, Mark My Worlds? Like if I said Mark My Islands, you’d be like -

 

[Krugerrand]

It is!
View attachment 506469

Candles good? Teapots bad? Why?

 

[MC3OZ]

On thing I got from the 3rd Row Podcast was how important GF Shanghai and GF Berlin are to unwinding the wave and in turn what a big difference that will make to many aspect of Tesla.
We knew that already but nothing like hearing it direct from Elon...
Assuming GF Berlin can eventually make RHD cars for the UK, then the primary non-US production out of Fremont is RHD cars, for the Asia Pacific, Hong King, Australia, NZ, etc....
These RHD Asia Pacific cars are likely to be the remaining instance where Tesla needs to pay suppliers before the customer takes delivery... They are smaller volumes .....
Local production and unwinding the wave, means time from order to delivery should be shortened in all markets (for Model 3/Y), it should be a smoother process with better flow and overall a better customer experience...
As well inventory and the mix of inventory can probably be more finely tuned, they don't need to stock pile as many cars when time from Order -> Manufacture -> Delivery is short.

 

[Krugerrand]

With all these traditional OEMs struggling to get enough batteries, there is always this nagging question in my head: how did Tesla, in such a crowded and competitive market, get their hands on production capacity at LG and CATL for at least 150,000 Chinese Model 3s per year (and more when the Model Y plant comes online)? How are they snatching those batteries away in front of VW, Mercedes, KIA/Hyundai, who are all extremely battery constraint, and not pay their weight in gold? Does anybody have an explanation?

Elon Musk.