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.
Tesla will spend 500M over 5 years for Dojo

So this means Elon will be spending 5x on Nvidia/AMD over Dojo this year.

Well they are both production limited. But Tesla is paying trough the teeth for NVidia and are getting Dojo chips at cost.
 
Tesla will spend 500M over 5 years for Dojo

So this means Elon will be spending 5x on Nvidia/AMD over Dojo this year.

$500M on this project, don’t think we can assume $0 on other locations…
 
I stopped at a 40 stall location in Dillion SC and I was the only car charging. This $1.4m investment will take a few years to achieve a payback.
However, the Supercharger at the Florida Mall in Orlando has 16 Superchargers that appear to have a line all the time. I bet the payback is less than a year.
Agree with that analysis.

But that is one way to look at it - short sighted view. The other way to look at it, the 16 charger location that was full with cars waiting perhaps turned away many potential customers who were horrified (quite rightly) at the prospect of waiting to charge, and the dozens of empty chargers (especially at places like Bucees) helped reduce the mental friction for ICE owners to put in an order.

The bad publicity of waiting cars vs. the very positive publicity of cars coming in, plugging into one of many empty chargers and leaving in 15 -20 minutes in a frictionless smooth operation, far outweighs any loss or profit numbers in installing & operating a supercharger - just my guess.
 
Last edited:
Elon Musk on X: "The governor is correct that this is a Dojo Supercomputer, but $500M, while obviously a large sum of money, is only equivalent to a 10k H100 system from Nvidia. Tesla will spend more than that on Nvidia hardware this year. The table stakes for being competitive in AI are at least several billion dollars per year at this point. " /X​

Lol, the way that Elon pumps Nvidia, I sure hope he owns a mitt-full of LEAPS on NVDA shares, so he can afford the taxes on his TSLA share options...

Cheers to the LEAPers! :p
Nah, he's just making sure NVidia doesn't stick Tesla at the back of the allocation queue for the newest chips. Tesla is walking a fine line with their Dojo project. If Jensen thinks Tesla will be a competitor to NVidia, you can kiss those sweet H100's goodbye!
 
Nah, he's just making sure NVidia doesn't stick Tesla at the back of the allocation queue for the newest chips. Tesla is walking a fine line with their Dojo project. If Jensen thinks Tesla will be a competitor to NVidia, you can kiss those sweet H100's goodbye!
Meh, AMD's MI300 is just as fast at 1/3rd the price. Just require a little bit of leg work for IT to get it running. I am glad now there's a competitor so people are not hostage to those price gauging Ngreedias.
 
Indeed, I hope Elon is sandbagging the DOJO hardware business until such time as we're ready to enter the market as competitors in this high-margin business.
This is what I was wondering. Based on how Elon described Dojo on Wednesday he made it sound like it just wasn’t working out. No updated charts about compute growth in the deck (as far as I know). Basically said Tesla would have to be super-lucky to succeed with Dojo.

Then two days later he comes out and says Tesla’s investing $500M in a Dojo compute center?

How do we interpret that other than Elon is downplaying Dojo? It must be doing something right if they’re going to spend half a billion on it in NY.

Edit: someone pointed out he may be downplaying it to avoid losing access to NVidia/AMD chips. That’s a good theory.
 
This is what I was wondering. Based on how Elon described Dojo on Wednesday he made it sound like it just wasn’t working out. No updated charts about compute growth in the deck (as far as I know). Basically said Tesla would have to be super-lucky to succeed with Dojo.

Then two days later he comes out and says Tesla’s investing $500M in a Dojo compute center?

How do we interpret that other than Elon is downplaying Dojo? It must be doing something right if they’re going to spend half a billion on it in NY.
Whenever Elon thinks, you always need to apply the possibility of him thinking in "production terms", "cost terms", or "productivity/utility terms".

1. Is Elon saying that production crunch of voltage regulators which will cause AI chip shortages in the up coming future which results in Dojo being the savior in this situation because it was made inhouse? Is this the high risk high reward he is referring to?

2. Is Elon saying that the cost of AI chips will just exponentially explode as more will see their potential usefulness and Dojo will pay dividend for Tesla if this happens?

3. Will training models be so big to do real useful things that only high speed interconnect platform like the Dojo design won't bottleneck progress?

To me Elon is saying that the high risk of developing something from scratch can potentially hedge against any of the problems he is predicting in his head to yield that high reward. So either Dojo needs to outperform the best AI chip companies have to offer or fully hedge against certain catastrophes which makes liklihood of success to be low but worth trying.
 
Last edited:
1706328597114.png
 
This is what I was wondering. Based on how Elon described Dojo on Wednesday he made it sound like it just wasn’t working out. No updated charts about compute growth in the deck (as far as I know). Basically said Tesla would have to be super-lucky to succeed with Dojo.

Then two days later he comes out and says Tesla’s investing $500M in a Dojo compute center?

How do we interpret that other than Elon is downplaying Dojo? It must be doing something right if they’re going to spend half a billion on it in NY.

Edit: someone pointed out he may be downplaying it to avoid losing access to NVidia/AMD chips. That’s a good theory.
Technically, Elon didn’t come out with the news. He just confirmed the news that came out about it.
 
  • Like
Reactions: johnm6875
Maybe because they didn't plan to simultaneously build manufacturing on three continents for the CT might have been a clue. 🤷‍♂️
Not at first, but I would not rule out eventual CT production, and certainly sales, outside the US.
From a style POV, I can see the CT selling very well in China, maybe the middle east. I expect quite a few people in Russia would like one. Europe's roads are too narrow, and pedestrian safety regs might rule us out.
Eventually, South America, India, Africa.

Tesla thinks super long term. Its very not-tesla to design a vehicle that will only ever ship to a single country. Besides, why bother shipping 8 cybertrucks to China for 'display' purposes, if you don't have any intention of selling any there?

Having said all that, I do wish they would have more urgency with the semi. There is a product that they could sell globally tomorrow, if they get going on megacharger installs.
 
-Chinese OEMs are the ones Elon talks of as the best competition;
-Chinese direct investment in electrical infrastructure, hydroelectrics, photovoltaics, Electric vehicles ( BYD -started with building electric busses in Brazil in 2015);
-Chinese construction of non-Chinese brands has resulted in faster market success that is largely ignored, e.g.. MG, Volvo, London taxi;
-Chinese battery and photovoltaics have grown to global dominance;
- Chinese infrastructure leaders such as State Grid are almost always led by local and/or non-Chinese expats, with the Chinese executives usually less visible. For example BYD in Brazil has a Brazilian leading executive whose face is often seen in news specials and interviews.
- The Chinese, as the largest global producer of renewables and autos continue to do very extensive effort to put a purely local face around the world, with such strategies as purely local production by locally consoled companies. one example os CAOA in Brazil unknown outside Mercosur but producer of Hyundai and Chery domestically and regular recipient of highest quality awards, so Chery has followed Hyundai as high sellers and consumer favorites.

I am mentioning more brazil for obvious reasons. Geely, though, has been highly successful with Volvo, which is Chinese-owned, small minority Swedish public.

The US and now EU are beginning to put trade barriers to Chinese development, but entities such as VAG, TSLA, MB, BMW and GM are all having excellent results there, so more efforts are happening to follow the pattern of arrangements such as that of CAOA in Brazil, to decrease susceptibility of political impediments.
One highly successful example of that is Canadian Solar:
It is strange that the only informed people who do NOT see the Chinese clearly seem to be North Americans. With ~75% of Amazon sales in US being Chinese origin we should begin to see what has happened.
In much of Africa, South America, large parts of Europe, most of Asia the problem is finding NON-Chinese high quality and cost effective products including cars and essentially everything renewable. Try finding an electric bus around the world with no major Chinese components.

For every one of us who are TSLA investors all we really need to do is look at Tesla Shanghai and Tesla sourcing of major components. That will establish that, at a minimum BYD and CATL are globally powerful competitors. Then, just ask Elon for his views, as he expressed them this week.

Each time we dismiss the most advanced and agile competitors we miss just how TSLA is one of very few to actually exceed them in some ways.

Finally, we credit TSLA with Giapress, jsutififiably. We also credit IDRA, justifiably. Then, most of us ignore that it was only Chinese ingenuity and determination that helped Tesla to accomplish that feat. A typically modern Chinese approach is to be found in IDRA and Volvo.

Perhaps it is natural to ignore just how important Chinese talent and innovation are to our own investments in TSLA.
In addition to your excellent post, I wanted to add some perspective from an Australian EV market as it has some interesting characteristics that may make it a bellwether or stepping stone for wider Chinese vehicle acceptance in western countries:

Some background characteristics of the Australian market:
  • Australia is wealthier by median wealth than virtually any other country - meaning a large portion of the population has the financial resources to purchase a wide variety of vehicles.
  • 1706352644590.png
  • It has a similar western standard / ideology to EU/US - although somewhat more Asia focused due to location / recent migration
  • It is a laggard in EV adoption as there are limited emissions standards (7% of new vehicle sales in 2023, up from 3.5% in 2022), limiting supply when manufacturers can send EVs elsewhere and get the credits
  • The rollout of non Tesla charging infrastructure is relatively slow, but better than in the US
  • There are virtually no import tariffs on vehicles to Australia and no domestic manufacturing competition
    • All vehicles therefore include transport costs to Australia, removing this as a price advantage vs domestic production seen in EU/US - Transport from Asia is cheaper than from EU/US
The below video from Fully Charged details how Chinese EVs entered the market (summary below):
  • China is #3 exporter of vehicles to Australia (behind Thailand and Japan)
  • BYD entered the Australian market 18 months ago and is now in the top 3 EV sales (Tesla first), with most of the remaining (ex Tesla) top 10 models also coming from China. Tesla is still the runaway lead.
  • 1706353533247.png
  • Consumer technology in Chinese EVs is impressive (compared to legacy auto), but maybe a little too tech focused for western audiences which focus more on driving experience
  • Legacy auto can't match the tech or quality of Chinese EVs at the price point
    • Chinese EV's can be as little as 1/3rd the price of legacy EVs - you can forgive a lot of issues for that sort of saving. Price difference can break through brand loyalty
    • It also unlocks a larger portion of the market that can afford EVs
  • Chinese EVs provide variety of makes and models for consumer interest
    • There are a lot of new models still to make it to Australia (pickups, GAC Aion with high power and fast charging to help long distance driving - cherry and zeeker are also rumoured to arrive in 2024, with Nio and Radar working through trademark issues)
  • Barriers to entry:
    • Homologation to Australian design rules is an impediment to Chinese manufacturers (but not more so than other brands)
    • Brand and perception
    • After sales support

Other comments:
  • Australia's auto market is relatively small (c.1.2m vehicles in 2023), this means that the manufacturing base problem is less of an issue - China sold c.8m pure EVs in 2023, so suppling Australia would be a rounding error and not require significant growth / restructuring to supply (Australia + China) market
    • As their manufacturing base increases to meet domestic needs, the % increase in their manufacturing capacity to meet meaningful market penetration of other export markets (EU/US) will decrease
  • I'm only in my early 40's and have seen Australian adoption of "rubbish" Korean runabouts that were laughed at in the 90s to them now placing p3 and p5 in the market. Only one US and one EU OEM make the cut.
  • 1706356795572.png
  • BYD are building EV manufacturing in Europe already
  • From our latest info there is still 2-3 years before Tesla NG will hit volume from GigaMexico, if Chinese manufacturers are planning expansion now, they may not be far behind that timeline
I'll be watching closely what happens to EV market mix and competitor pricing in Australia over the next couple of years. This doesn't mean I think Tesla will fail due to Chinese dominance, but it would be foolish not to be watching them closely and copying any advantage we can find from them.