Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

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

This site may earn commission on affiliate links.
Also, the question needs to be asked as to how come X and Xai had the physical space to absorb tens of thousands of GPUs, whereas Tesla didn’t? Did Xai have any giant empty data centers sitting around? Don’t think so. So whatever solution they came up with for using the GPUs (3rd party data centre space?) why couldn’t Tesla do the same in its most important current goal of achieving FSD.

Why wasn’t Tesla ready to receive the order they had made? What failure of management took place that a half billion dollar order of an incredibly valuable and scarce resource had nowhere to be put to use despite a long lead time? The southern extension is still months away from completion. Was there not room that could have been prepared elsewhere in one of americas biggest buildings?
You can't just plop them anywhere. You have to have lots of cooling and power available. (And networking.)

As far as why Tesla didn't have room? I seem to recall someone was fired over the GigaTexas expansion, that was to house the datacenter, not being completed on time. So, the plan was for space to be available, but something happened that caused it to not be ready. We don't know why that is. (And I'm not sure Tesla even needs to be sharing that level of detail with the public or shareholders.)
 
At the rate things are moving, Tesla will not have a first mover advantage. The incumbent truck manufacturers are developing their own cell supply to scale up EV manufacturing.

On the other hand:

The incumbant manufacturer Nissan's Leaf arrived on the market about a year and a half before the Model S.

The incumbant manufacturer Chevy's Bolt beat the Model 3 to market by about 6 months.

Speculating: Rivian and Ford both beat the Cybertruck to market...but:
  • I think it is very likely that the Cybertruck will outsell the Rivian R1T this year
  • I think it is quite possible that the Cybertruck will outsell the Rivian R1T and R1S COMBINED this year
  • I'm not sure what Ford's plans this year are, but they sold fewer than 25,000 F-150 Lightnings in 2023. Based on that, I think it is quite possible that the Cybertruck will also outsell the F150 Lightning this year.
  • 2025: what are the chances the Cybertruck could outsell all the other electric pickup trucks combined in the US next year?

Sometimes, the product itself, the company's production plans, and the company's manufacturing capabilities matter more than first mover advantage.

No guarantees either way of course...but the only way to know is to watch and see :).


*Edited to add the pickup truck data
 
Last edited:
On the other hand:

The incumbant manufacturer Nissan's Leaf arrived on the market about a year and a half before the Model S.

The incumbant manufacturer Chevy's Bolt beat the Model 3 to market by about 6 months.

Sometimes, the product itself, the company's production plans, and the company's manufacturing capabilities matter more than first mover advantage.

No guarantees either way of course...but the only way to know is to watch and see :).
Not a valid comparison. The leaf was a short range city car. The model S was the first long range EV of the modern era. Tesla was a first mover in the long range BEV market and held a cost advantage for a long time.

No one currently makes electric semis at scale but the incumbents are gearing up to do so. The longer the semi is delayed, the less likely it is to have the scale, cost, charging and brand advantage that Tesla cars enjoyed in the early days.


 
There are now so many posters like this, who do nothing but post negative-spun FUD, that its ridiculous.

I just dont get why people spend SO MUCH TIME on an investment forum dedicated to a company they hate so much.
110% agree!

...Leading to the logical question (to the powers that be?) ... why are they still here? Especially since actual valuable members (who cares if they are a bit spicy) seem to have been turfed. What is the strategy ... kill the forum?

I would really like to understand. M
 
I just find it silly that Elon had to explain this. You just need to look at the finance statements and the original tweet. Cap EX is running about $10B year. It can't be all Nvidia chips. It was well understood he meant the hardware in the cars as well as internal expenses.

 
Last edited:
110% agree!

...Leading to the logical question (to the powers that be?) ... why are they still here? Especially since actual valuable members (who cares if they are a bit spicy) seem to have been turfed. What is the strategy ... kill the forum?

I would really like to understand. M
They won't say, but it can't be good. I'm fading slowly. Surprised I'm not banned yet.
We became too powerful. Plus, I was getting too good at predicting the stock and some people didn't like it.

Good news, this is all very, very bullish as they dig in their heels even more.
 
Not a valid comparison. The leaf was a short range city car. The model S was the first long range EV of the modern era. Tesla was a first mover in the long range BEV market and held a cost advantage for a long time.

No one currently makes electric semis at scale but the incumbents are gearing up to do so. The longer the semi is delayed, the less likely it is to have the scale, cost, charging and brand advantage that Tesla cars enjoyed in the early days.



There are many reasons why things weren't directly comparable, and that was part of my point and why I said that the product, the company's goals, and capabilities matter IN ADDITION to things like "first mover advantage."

Of course the Nissan Leaf was a less capable product, it was also less than half the price. But there were plenty of comparisons at the time. And, Nissan's production goals were never as lofty as Tesla's. Similar to the Bolt -- a bit cheaper than the 3, lower performance, but range in-between the SR and LR Tesla...but Chevy never really planned the production numbers that Tesla aimed for. Of course there were people who predicted victory for the Bolt since it hit the market first...but in the end other factors mattered more.

In this case, from your linked article:

"Total investment is expected to be in the range of $2-3 billion for the 21-gigawatt hour (GWh) factory."

I believe that cell capacity is roughly half of what Giga Nevada currently produces...while Tesla is also buying even more cells from other manufacturers, plus ramping and improving 4680 production, plus Tesla's suppliers are also building 4680 factories and also ramping other cell types.

So, this 21 GWh news is 3 incumbent manufacturers combined, and I didn't see the target factory completion date or the expected ramp to full production.

Ballparking a semi battery between 500 kWh and 1000 kWh, 21 GWh is enough capacity for these three companies to produce a total of 21,000-42,000 trucks combined annually.

A sizeable number to be sure...and there are probably additional industry plans as well...but I'll leave it to others to speculate on when those big numbers might be achieved, and what the Tesla Semi production might be by then.
 
I just find it silly that Elon had to explain this. You just need to look at the finance statements and the original tweet. Cap EX is running about $10B year. It can't be all Nvidia chips. It was well understood he meant the hardware in the cars as well as internal expenses.

Well, considering the "need to know" policy when it comes to executive decisions has failed majestically, the next best thing is for him to be on top of these rumors. Don't want to explain? Then don't let anything leaked. To be clear, I already knew the reason before his tweets, but I'd prefer them over silence.
 
I just find it silly that Elon had to explain this. You just need to look at the finance statements and the original tweet. Cap EX is running about $10B year. It can't be all Nvidia chips. It was well understood he meant the hardware in the cars as well as internal expenses.


From Elon's tweet:

"My current best guess for Nvidia purchases by Tesla are $3B to $4B this year."

So, $4B sounds like another absolute, binding "promise" for critics to latch onto...
 
This is interesting:

Training compute for Tesla is relatively small compared to inference compute, as the latter scales linearly with size of fleet. Perhaps the best way to think about it is in terms of power consumption. When the Tesla fleet reaches 100M vehicles, peak power consumption of AI hardware in cars will be ~100GW. Training power consumption is probably <5GW. These are very rough guesses. Obviously, 5GW of AI training compute is enormous by current standards, but is only about 5% of total Tesla AI compute.There is a path for Dojo to exceed Nvidia. It is a long shot, as I’ve said before, but success is one of the possible outcomes.
100m Tesla vehicles HAS to be confirmation of a model 2, alongside CT,RT,3,X,S,Y,Semi.
Sure, Elon is looking ahead, but even 10 years ahead, to get 100m you need 10m production a year right? I'm surprised there is not more raising of eyebrows about this.
I think this is evidence that 8/8 might reveal a robotaxi in development AND a same-production-line model 2 you will be able to get 2025. It makes perfect sense for Tesla to do it. The fact that elon doesn't tweet about it just means its not HIS focus. I bet tom zhou etc are very focused on the unsexy, but likely very popular compact/cheap model 2.
 
Yup. Dumb. Tesla didn’t have the space for them.


"We placed an order for NVidia chips to use in our mission critical project. Then we sat on our hands for months, and when they were about to ship, we suddenly noticed we forgot to have a place to install them in."

Amazing how xAI, founded in March 2023, had the space.
 
Quoting the below comments, as they seem to have been missed and need repeating.

You can't just plop them anywhere. You have to have lots of cooling and power available. (And networking.)

As far as why Tesla didn't have room? I seem to recall someone was fired over the GigaTexas expansion, that was to house the datacenter, not being completed on time. So, the plan was for space to be available, but something happened that caused it to not be ready. We don't know why that is. (And I'm not sure Tesla even needs to be sharing that level of detail with the public or shareholders.)
If you have worked in supply chain/logistics this kind of thing goes on constantly, sometimes even between competitors. Everyone is trying to manage cashflow and physical space and *sugar* happens. Suppliers will usually be the go between to negotiate something beneficial to everyone.
Exactly... and this is why the CNBC article is SO wrong. It is because Elon Musk owns these other companies [that can use Nvidia's chips] that Tesla is not calling up Nvidia to say "ummm, remember those chips we ordered? Well, we're not really ready for them yet, go figure out something else to do with them" at which point Nvidia would sell them to someone else. Elon simply said to Nvidia "Tesla isn't ready for them yet, but I know another customer who is." Anyone trying to cast this situation in a negative light misunderstands the relationship between Elon and Jason. Tesla was one of the first high-profile, non-Bitcoin-mining customers of Nvidia's GPUs that was definitely not using them for PC video cards. Tesla selected Nvidia when they decided to move away from Mobileye in 2016. (Tesla already was a long-time customer of Nvidia for the Tegra processors in the MCU1 in 2011) Nvidia remains grateful for the relationship I'm sure adn has no problem shipping the chips to whatever address Elon supplies.

Imagine if General Motors decided to delay/cancel an order. They do not have some other company that needs AI processors. In fact the supply chain problems during COVID [which were nothing to do with the virus] arose specifically because auto manufacturers cancelled a bunch of orders, and then the chip fabs started favouring other types of customers that weren't screwing them around - and then when the auto companies needed chips again, it took them a while to get inventories back to the right levels. Nvidia is happy that Tesla has not been screwing them around cancelling orders.
X was expecting delivery in January. The memo about swapping delivery dates was from December. X was already building out their data center whereas Austin wasn't even close to ready.
12k H100 = 12MW of power (700W per plus balance of plant). Have you seen the additional conduit and cooling towers getting installed? This is not something one just drops into an exsiting building.

The piece is dumb. All that happened is the delivery dates were swapped.
Elon's investment number likely includes vehicle inference (current production and new chip versions), data centers, Dojo, staff, and H100 purchases. Nvidia only has visibility on one of those buckets.

In short:
1) All Elon did was swap delivery dates for NVidia hardware between two companies. Obviously, not theft. The chips were getting made and paid for one way or another...if one company was ready for them sooner, it makes sense to swap delivery dates instead of storing the hardware in a warehouse or trying to cancel/delay the order.

2) This sort of swapping happens often in industry, even between competitors, because it keeps the suppliers happy and stable.

3) You can't just plop the NVidia hardware anywhere and make it work, even if you have empty space somewhere else in the same building. The cluster needs all sorts of electricity and data lines and cooling systems, plus all the shipping and installation logistics. So, if there's a delay in building that infrastructure, the chips can't be used for some time unless an alternate home is found. And this happens often in industry, even between competitors.

4) At least one commenter (quoted above) remembers there being a delay in the construction that got somebody fired. So...perhaps there was a problem or some problem that could have been avoided. Many companies have avoidable problems that eventually get resolved...that's not really news (unless of course the headline can include "Tesla")

5) During COVID, we know that many traditional auto manufacturers cut production and cancelled chip orders. So, the chip manufacturers changed their plans and priorities to provide chips to other customers. When auto manufacturing came back up, the traditional auto manufacturers couldn't get the chips they wanted because they had been de-prioritized by the suppliers. The "troublesome" situation that is being brought up today is actually a good way to keep a supplier (Nvidia) happy and avoid this sort of issue.

If I was going to try to summarize I'd probably write:

"We (Tesla) placed an order for NVidia chips, probably many months or even years in advance due to high demand, because we knew we'd need them for our mission critical project. Then, we continued building our factory and the attached data-center to house and run those chips. A startup AI company, known as X and tied to Twitter/X, also ordered a whole bunch of chips. They happened to be behind us in line for chips...but being tied to Twitter, they might have already had some data center space available. Our factory/data center construction didn't go as quickly as we had hoped, so we (Tesla) swapped line spots with xAI so that they could get chips sooner, and we'd get our chips when we can actually install them instead of having to store them for a few months. Both companies get what they need when the need it, and NVidia as a supplier stays happy too.

Amazing how these things can work out positively for everybody sometimes, and certain media sources will still try to spin it as a problem!"
 
Last edited:
This is interesting:


100m Tesla vehicles HAS to be confirmation of a model 2, alongside CT,RT,3,X,S,Y,Semi.
Sure, Elon is looking ahead, but even 10 years ahead, to get 100m you need 10m production a year right? I'm surprised there is not more raising of eyebrows about this.
I think this is evidence that 8/8 might reveal a robotaxi in development AND a same-production-line model 2 you will be able to get 2025. It makes perfect sense for Tesla to do it. The fact that elon doesn't tweet about it just means its not HIS focus. I bet tom zhou etc are very focused on the unsexy, but likely very popular compact/cheap model 2.
This is another area people consider Elon a liar. Him saying When Tesla gets to 100m cars doesn't mean anything. He's talking about hypotheticals, it could be 10 years or 50 years. In 10 years if Tesla doesn't have 100m total cars sold, he didn't lie, but he'll be cited by someone (not you) that he promised 100m cars or a M2 coming soon. It's the same with his 20m cars a year by 2030. It wasn't any sort of guarantee.

This isn't directed at you, just thought it wrapped up that "liar" conversation up.
 
Not a valid comparison. The leaf was a short range city car. The model S was the first long range EV of the modern era. Tesla was a first mover in the long range BEV market and held a cost advantage for a long time.

No one currently makes electric semis at scale but the incumbents are gearing up to do so. The longer the semi is delayed, the less likely it is to have the scale, cost, charging and brand advantage that Tesla cars enjoyed in the early days.



Mmm... I'm not so sure on the "valid" part... there were lots of folks for whom the criteria for a decent daily driver was "commute distance", and the Leaf was capable of that, The fact it looked like a bug, was not marketed well, and was not engineered well are the more likely reasons for tepid adoption, and it's getting subsequently clobbered by the 3 and Y.

But I'll go ahead give you that. The Bolt, however, that I mentioned earlier, did have the range to go along with it's first-mover advantage. And it also got blown away.

The Semi "competition" might be the next round of "Tesla Killers" that has given this forum so much entertainment over the last decade.
 
On the Semi, Tesla definitely had first-mover advantage for a while, but has given it up. I know people working for large consultancies who assist companies in green transportation / logistics initiatives. None of them are considering Tesla in their packages because Tesla isn't really delivering -- they've shipped less than 100. Other major commercial truck companies are delivering electric, hybrid and alternative fuel vehicles today.

That said, I suspect the Tesla ones will win in the end. They're more forward-looking.
 
On the Semi, Tesla definitely had first-mover advantage for a while, but has given it up. I know people working for large consultancies who assist companies in green transportation / logistics initiatives. None of them are considering Tesla in their packages because Tesla isn't really delivering -- they've shipped less than 100. Other major commercial truck companies are delivering electric, hybrid and alternative fuel vehicles today.

That said, I suspect the Tesla ones will win in the end. They're more forward-looking.

What other BEV Semis of similar range are being delivered in significant volume today?