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.
Is AMD Zen the GPU they advertised as PS5 equivalent when Plaid came out? I honestly don't know.
Seems to be a lot of confusion on this.

Model 3/Y now has the Zen+ 4 core/8 thread processor WITH integrated Vega graphics (PC equivalent is an AMD 3400g processor made from Global Foundries 12nm fab)

Model S/X has the Zen+ 4core/8thread processor with integrated vega graphics + dedicated RDNA 2 GPU (PC equivalent of a 3400g processor + 6700XT, the 6700XT is made from TSMC's 7nm fab).

The difference in graphical power between the Model 3/Y Zen based APU vs the one on the S/X is about 10x. However, the integrated Vega graphics is about 2.5x faster than what was in the Atom

For those who worry about chip shortages, there are PLENTY of 12nm chips. Frequently we see discounts on these processors in the OEM market. The 7nm graphic processor has more of a supply issue. However given that 10s of millions are made per year, you don't need to worry about it from a S/X volume standpoint.

More technical clarifications.

FSD computer is based on a 14nm node fabbed from samsung
DOJO is a 7nm custom chip on wafer technology fabbed from TSMC
 
Last edited:
Point of clarification, if you please:

With respect to the ”Tesla’s App” conjecture - what kind of apps might it be appropriate for a Tesla, as opposed to a platform provider like Apple or Alphabet, to provide? Would it be likely that such apps would…could…should be available other than on vehicles’ screens? For a typical user, would it be easy to “parse” between one’s set of “phone apps” and “Tesla apps”?

A little bit separately, do any have well-founded opinions as to whether the revenues & margins Apple obtains through its universe of apps be appropriate to consider for Tesla?
Why not just make an app store compatible with android? I believe that is what Microsoft is doing on windows.
 
  • Like
Reactions: Surfer of Life
OT weekend thinking about chips (the kind that go tick/tock) and maybe this should go into the Engineering thread (tag me to discuss in that thread how abouts...)

TL;DR - For Tesla, as a business, I think the need for having their own fab(s) is at or near the tipping point, as lack of scale is the show stopper and that issue is about to be eclipsed. Fabs are a huge capital investment and the timeline is 2+ years (even for Elon) once you have your chip roadmap and architectures nailed to build the machines that make the chips. Dojo would be a good Trojan Horse or a way to 'cut your teeth' on making near infinitely valuable compute that would go into the droid. If Tesla's ideal chips could come in a smallish set of types/classifications then it is a forgone conclusion they should pursue fab(s).

(Some high level tech PM weekend hand waving below...)

Today, Tesla can flash any chip in the vehicle. If they can do that, why not just send and receive constantly from a set of centralized compute cores to those vehicle systems and sub-systems to remove the localized compute? For one, they need to get off of the CANbus which is licensed and static. Move to their own deterministic communication method. A crazy amount of work, but not intractable.

I think an ideal draft design would be to have a set of two massive compute chips in lock-step, each with fully redundant power, communication and physically separated which contain all necessary cores for independent compute and networking layers, memory separation as well. This might not be the next step for Tesla depending on what the current designs include. Based on that, I'd assume you'd build a roadmap to get to the final design as jumping from current design to the most efficient design, would most likely cause build bottlenecks. I could see them developing more like 5 different chips based on classifications in the shorter term, as a stepping stone, and eventually moving to just one.

As some of us know, TSMC is one if not the most profitable company on the planet at making chips. They do this at scale and are pushing the envelope constantly so they can charge a pretty penny for their products over that products lifetime. They can also move customers to new products sooner than they'd want and charge another premium to help build the runtime stack, service stack, app platform, support stack, dev tools...etc. It is immensely profitable and will only become more and more so overtime.

It isn't, however, lost on me how powerful the current AMD Ryzen chip is though and that could be a good central compute for the time being. However, it might not be great at running Tesla's Linux as efficiently as a home grown architecture. And also, might not be robust enough to handle the compute needs for a fraction of the other ECUs compute needs.

Realize that Dojo and the Inference V4 chip will share similar components to generic generalized CPU compute. All custom AI chips require some amount of generalized compute. The ability to actually manufacture these chips would possibly be the next step, but Tesla might be utilizing some 3rd party patents for some parts of the Dojo or inference chip. So the next step would be to design a patent-free chip or establish a perpetuity free use license for critical IP. After this is known good and in parallel, learn to manufacture while iterating on a patent free or license-able fab that can meet your scale and classification needs.

Tesla has the chip design folks, but they'd need to either, hire fab design folks, buy IP for fab design or just buy a fab and dig in. Potential cost savings are staggering **IF** you knew you were going to scale to needing at least 1 billion chips/yr total and less than 10 fabs/lines. A "one chip to rule them all" design might be best and then just de-pop or fuse what you don't need for different needs. The ability for a single chip design, single footprint, single OS and stack would be a huge benefit across the board **IF** you had a high enough volume to justify the capital investment. Just a single interface, layer, deterministic design for networking would be enough to justify IMO as the gains in efficiency would be staggering. To not have to deal with any legacy interfaces and to know that delivery of messages throughout the car were going to arrive in a very tight tolerance makes the overall communication infrastructure quite simple relatively speaking.

Due to their specific, yet somewhat generic needs, the amount of chip designs they'd need for all their announced products is nearing 100 million/year (my estimate as each car, powerwall, charger, supercharger, eventually the droid..etc uses several chips per product)

This starts, for me, with the canonical question, how many of these chips are you going to need?

And ends with, let's say you have all the ideal chips you need, what then?

There's so much there to unpack...but hey, it's the weekend right?

Q: How many compute nodes are in a Tesla today? What does the BOM say?
A: Let's call this variable X
Q: Do all these nodes perform the same level of compute?
A: For simplicity sake, I might break this into classifications. Basic Linux OS runtime, rudimentary computations, specialized critical multi-layer, AI NN training, AI NN inference.
Q: Are there other needs outside of the car that the car classification of chips isn't covering?
A: I think they are covered
Q: Does every chip classification need a networking layer?
A: Totally, this is how we get to ideal communication in the ecosystem
Q: So, how many chips are we talking about?
A: Let's say there's a need for 100 million in 2025 to 1 Billion by 2030
Q: What is stopping Tesla from pursuing this?
A: Engineering first principles logic is NOT as simplistic as I've outlined so the answer is nuanced.
Personally I don't think there will be a recession, but hypothetically if there was a recession at some stage, that might be a great opportunity to buy a company that owns several fabs for a good price.

Or a some time in the future Tesla and a company in the chip business might see synergies in merging their business and perhaps creating new product lines.

For now the chips shortage seems to be easing, so there is no rush.

But for Tesla and SpaceX chips are important parts of their business just as vital as batteries.

I prefer the general approach as per Battery Day have a long term vision and establish a series of targets in terms of supply, price and performance, Acquire the the technology/companies necessary to achieve the vision. Prove all concepts then slowly incorporate in house supply into the inventories.

I see solar in similar terms and I'll post in the engineering thread one day, for starters the light spectrum extends for Ultraviolet to Infrared. New technologies are discovering how to extend the range of light frequencies that are tapped. Some of this has implications for the performance of solar on cloudy days. Research is finding that things like replacing silver in solar panels with copper can increase efficiency, and lower prices.


Yes, solar is a long shot, but it is part of the business and the mission, also research is far from fully tapped out, there are many ways of improving efficiency and lowering costs, a lot of which are still in the research/start up phase. Again a lot of the R&D and technology that would be needed can simply be purchased. We know Tesla are good at building and designing factories, but again no need to rush.
 
It's more like $1B+ of R&D per quarter just to stay in the process game (hard-core material science) and a (big) FAB runs at $10B+ and needs to be upgraded every few years.
A smaller plant will be cheaper, but unless they want to reinvent industries like isolation foundation construction, clean rooms, online purification of gases/liquids, chemical handling+reprocessing, wafer production (?), um/nm precision positioning, optics, laser sources, substrate production and chip testing methods it won't be significantly cheaper..

Might just as well use this money to buy themselves to the front of the line at whoever is the current leader on the target process.

But besides that, the current shortage seems to be legacy car controllers built on 'ancient' process technology for which no new production equipment will be built. And they aren't redesigned for current processes because it's extremely low margin and there are high costs for redesign, validation & certification. So Tesla would have to finance those costs for every chip they want to insource.
All very good points and that does not even cover the IP issues which prevent new entrants. There is a reason there are 3-4 biggies.
 
  • Like
Reactions: transpondster
Thank you.

At what number would make sense for Tesla to have a FAB?

Is there a limit to the amount of stuff that would use a chip, any chip?

The need obviously needs to be there, if they decide to have a FAB.
The answer to this is never.

To put things in perspective

TSMC total property/Plant cost : 71 BILLION dollars
TSMC Quarterly Revenue: 15 Billion


Tesla Total property/plant cost: 18.8 billion
Tesla quarterly revenue: 15.3 billion


You tell me what Tesla should focus building more of. Fabs are gigantic money pits and building them from scratch is nonsensical. Also Ferragu comparing Tesla to TSMC is nonsense. Tesla's revenue dollar/capex is on a whole new level compared to TSMC.
 
Seems to be a lot of confusion on this.

Model 3/Y now has the Zen+ 4 core/8 thread processor WITH integrated Vega graphics (PC equivalent is an AMD 3400g processor made from Global Foundries 12nm fab)

Model S/X has the Zen+ 4core/8thread processor with integrated vega graphics + dedicated RDNA 2 GPU (PC equivalent of a 3400g processor + 6700XT, the 6700XT is made from TSMC's 7nm fab).

The difference in graphical power between the Model 3/Y Zen based APU vs the one on the S/X is about 10x. However, the integrated Vega graphics is about 2.5x faster than what was in the Atom

For those who worry about chip shortages, there are PLENTY of 12nm chips. Frequently we see discounts on these processors in the OEM market. The 7nm graphic processor has more of a supply issue. However given that 10s of millions are made per year, you don't need to worry about it from a S/X volume standpoint.

More technical clarifications.

FSD computer is based on a 14nm node fabbed from samsung
DOJO is a 7nm custom chip on wafer technology fabbed from TSMC
I mean if this is true that's completely crap for 3/Y owners, only S/X owners will be able to play The Witcher 3.
 
  • Informative
Reactions: Surfer of Life
Thank you.

At what number would make sense for Tesla to have a FAB?

Is there a limit to the amount of stuff that would use a chip, any chip?

The need obviously needs to be there, if they decide to have a FAB.
If Apple, the largest single consumer of chips on Earth and the world's most valuable company by market cap, doesn't need to have a fab, why would Tesla ever need one?
 
I mean if this is true that's completely crap for 3/Y owners, only S/X owners will be able to play The Witcher 3.
They never advertised Witcher 3 for the 3/Y. However the 3/Y screen is a lower resolution screen and the integrated graphics is capable of playing witcher 3 in 720p high.

However 3/Y most likely are no equipped with proper storage for games like Witcher 3 and Cyberpunk anyways as those are 20-30gigs each.

.
 
I wonder if they will ever offer AMD Zen retrofits to owners of cars with Intel Atom. If not, I'll never be able to play The Witcher 3 on my car.

Last I read there was an internal letter specifically saying they would not.

IIRC another bit in the thread was someone having been told by an SC the new MCU was tied to the new 16v LiION battery that replaced the standard 12v, so retrofitting the MCU wouldn't be possible right now....but SCs sometimes give wrong info so unclear if that tying is accurate (nor am I clear how difficult retrofitting both the 16v battery and the MCU together would be).



Anyone have any sense on what code it would be? Might be time to dust off the old textbooks.

Flashlight app here I come.

Current MCU is running Linux AFAIK, and on an x86 CPU (in both MCU2 and MCU3 cars) so pretty standard code.

IIRC they're running straight up MAME for all the older atari type games for example.



Point of clarification, if you please:

With respect to the ”Tesla’s App” conjecture - what kind of apps might it be appropriate for a Tesla, as opposed to a platform provider like Apple or Alphabet, to provide? Would it be likely that such apps would…could…should be available other than on vehicles’ screens? For a typical user, would it be easy to “parse” between one’s set of “phone apps” and “Tesla apps”?

A little bit separately, do any have well-founded opinions as to whether the revenues & margins Apple obtains through its universe of apps be appropriate to consider for Tesla?


I think 30% is pretty standard margin, don't see why it wouldn't be appropriate.

That said- the number of apps that make sense to buy for the car (at least before FSD is L3 or higher) is vastly smaller than for a phone.... and the amount of local storage to hold such apps is also much smaller than most modern phones.

Not to mention the entire fleet is only 2ish million vehicles. So total revenue is not gonna be THAT exciting.

Even on phones, where there's a LOT more spend per user likely, average revenue is like $100 per device or something.... at 30% for Tesla that's 30ish million dollars.

It's not nothing- but it's pretty small versus Teslas revenue these days.


Anyway regarding type of apps-

Car-while-parked apps are pretty simple....games mostly I'd imagine, and likely mostly simpler ones (since the addressable market without the hefty S/X GPU is much greater- plus folks are more likely to want games they can play in small bunches at superchargers, rather than sit there for hours playing in their car).....

Media playing apps would be possible there too- but since the car has a web browser that enables both audio and video while parked not sure the value in a seperate app for most stuff there.



The big question would be how much Tesla is willing to allow 3rd party apps to do while the car is in motion. One of the long suggested reasons for resisting apple play was Tesla didn't want to give up control of the screen while the car was in motion. And given the deep integration of the nav with some driving features that's still an issue for things like a Google Maps, Waze, or Abetterrouteplanner type apps.

But those would certainly be among the most desired if Tesla enables integration with the driving systems.... plus maybe some other travel related things for finding hotels, restaurants, etc... (though honestly, not sure what much you get out of that on the car screen vs the phone)


Media apps probably fare a lot better here- since browser isn't an option, and having control on the car is a lot better than on the phone- doubtless some 3/Y owners at least would pay for a Sirius XM app, or audible, or many others.



Now, once FSD is capable of L3 or higher where you're allowed to be "paying attention" to a screen, that changes the game a fair bit and a lot more is possible.

But we're not there yet.



Why not just make an app store compatible with android? I believe that is what Microsoft is doing on windows.

That's....SORT of what MS is doing.

What they're actually doing is having Windows 11 run an Android virtual machine along side the actual Windows 11 OS.

The new Ryzen based MCUs would be able to do this (though depending on the full specs how well they can do it AND do everything else they need to at the same time might be a question)

Atom based MCUs would likely have a lot of trouble doing this.
 
They never advertised Witcher 3 for the 3/Y. However the 3/Y screen is a lower resolution screen and the integrated graphics is capable of playing witcher 3 in 720p high.

However 3/Y most likely are no equipped with proper storage for games like Witcher 3 and Cyberpunk anyways as those are 20-30gigs each.


yeah I haven't seen anyone publish the amount of onboard storage on either version of the MCU3 computer.

But that'll tell us a lot about Teslas plans for the next few years in the app store space.

App version of ABRP or an SXM media player would be pretty tiny.

AAA games are massive.
 
I wonder how long it will be before their financial situations get so bad that either Ford, or GM, or both, begin to be referred to as "embattled automaker Ford" "beleaguered car manufacturer GM" and so on. I suppose that gets answered by how much they are spending on advertising in those media. Once the advertising spends begin to shrink... perhaps the honesty will start flowing in the news reports.

It's too late for legacy media as well. As big auto declines, its advertising will shrink. Meanwhile more and more people are getting their news from quality YouTubers. Even advertisers will eventually realize that their spends on legacy media is worthless, and will start spending big on new media. I won't be shedding a single tear when that happens.
 
If Apple, the largest single consumer of chips on Earth and the world's most valuable company by market cap, doesn't need to have a fab, why would Tesla ever need one?

Because its an obvious attack vector for opponents of Tesla and its mission. Think Big Carbon (or 12 other disrupted industries). Who has a spare $71B laying around just to toss a **** into Telsa expansion plans? Who is funding the endless court challenges in Berlin? Who stands to gain by cornering the margin on chips? It all has one simple goal: to slow down the inevitable.

Ask yourself why the CIA is responsible for writing the U.S.A.'s annual strategic minerals report for Congress. National security is at stake if the U.S. can't get raw materials because a foreign country controls access.

For a history lesson, read about how in 1962 the CIA obtained titanium from the U.S.S.R. needed to build the A-12 Oxcart / RS-71 Blackbird:


Ironically, we know more about international spy rings than we do about corporate espionage. I guarantee its happening (and yes, some is state sponsored).
 
I’ve been saying this for years. Right up until they get in the car it’s pretty much the same old “I can buy a Camry for 20 grand less” spiel. Then they drive our model 3. It’s pretty much over after that. All of a sudden it becomes “worth it”.

One of the things I am doing on test drives now is to show them the ease of using a supercharger. I have them back in and hook up and then we walk over to get a coffee at Tim hortons. We come back, unhook, get in and go. It kind of gives a better context of the convenience of not going to a gas station and having to hold a gas nozzle. I also emphasize that the vast majority of charging they do will be at home in their driveway while they sleep.

Elon should put me on commission. /S.

Cheers. Stay safe all.

This is why Hertz will help creating more demand!
 
I’ve been saying this for years. Right up until they get in the car it’s pretty much the same old “I can buy a Camry for 20 grand less” spiel. Then they drive our model 3. It’s pretty much over after that. All of a sudden it becomes “worth it”.

One of the things I am doing on test drives now is to show them the ease of using a supercharger. I have them back in and hook up and then we walk over to get a coffee at Tim hortons. We come back, unhook, get in and go. It kind of gives a better context of the convenience of not going to a gas station and having to hold a gas nozzle. I also emphasize that the vast majority of charging they do will be at home in their driveway while they sleep.

Elon should put me on commission. /S.

Cheers. Stay safe all.
Similar vein: my dental hygienist (for the past 12 years) had always questioned why I got a Tesla four years ago, based on the Tesla/general EV FUD her husband told her (based on the Tesla/general EV FUD he had been told).

Last week, during a regular teeth cleaning session, my hygienist asked about my car, but the tone was totally different.

She proceeded to explain how her and her husband met some new friends a few weeks ago and went for a ride and test drive in their (the new friends) TM3.

Suddenly, her and her husband are looking at buying a TMY.
 
To put things in perspective

TSMC total property/Plant cost : 71 BILLION dollars
TSMC Quarterly Revenue: 15 Billion

Tesla Total property/plant cost: 18.8 billion
Tesla quarterly revenue: 15.3 billion


You tell me what Tesla should focus building more of. Fabs are gigantic money pits and building them from scratch is nonsensical. Also Ferragu comparing Tesla to TSMC is nonsense. Tesla's revenue dollar/capex is on a whole new level compared to TSMC.
Revenue comparison is not the best indicator, IMO.

Supporting TSMC's plant investment is that TSMC 2021Q3 Gross margin for the quarter was 51.3%, operating margin was 41.2%, and net profit margin was 37.7%, and those numbers are set, more or less, on a cost plus basis which they can since they have a captive group of customers with no real competitors in many area.

Tesla 2021Q4 operating margin was 14.7%.

Edit: Personally, I am don't see Tesla considering chip fab in 2022/2223, but depending on Teslabot it could be possibility for the far future.
 
Last edited:
Revenue comparison is not the best indicator, IMO.

Supporting TSMC's plant investment is that TSMC 2021Q3 Gross margin for the quarter was 51.3%, operating margin was 41.2%, and net profit margin was 37.7%, and those numbers are set, more or less, on a cost plus basis which they can set since they have a captive group of customers with no real competitors in many area.

Tesla 2021Q4 operating margin was 14.7%.
BUT fabs are only profitable with >95% uptime 24/7/365. And a huge money sink if you have slightly suboptimal processes...
TSMC is impressive, but i don't think Tesla would survive the learning curve.

Survivorship-bias + extreme conditions with chip shortage driving up prices are both at work here.
Just look at the fab manufacturers that has to close shop the last decade.