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

[Usain]

Which is why I am not sure I understand why one would want to build physical taxis when you can just own the software which everyone is forced to buy.

Its like Microsoft deciding the real dough is in making computers. Anyway, holding for the duration.

IBM did pretty well making both mainframe computers and the software that ran them. In 1967, it was worth $1.3 Trillion in 2020 dollars. In 2020, IBM ranked #5 in all time in inflation-adjusted market cap. Interestingly, that was ahead of #7 Microsoft.

Of course, Microsoft has passed IBM in inflation-adjusted market cap now.

I think Tesla will do just fine building physical taxis. It's a cost advantage nobody else will be able to match. And often for bleeding edge technology, you need to have full control of both hardware and software to make a really great product.

There is more than one way to build a great company.

 

[Iain]

no shingle lasts 50 years. the warranty is useless if you read it.

You may be right but ill be long dead by then so the next owner will have to deal with it

 

[loquitur]

You're right, he does make a lot of references to the power consumption of the hardware, and not as often in terms of the FLOPS or whatever computing metric people use nowadays.

In this case, however, he's talking about the cooling tower itself...so it makes sense to describe it in terms of the power it can dissipate, and therefore the power of the hardware it can cool. And from a "factory/building owner" perspective, running all the electrical power, and the cooling lines, etc. is the specification that is more at the forefront in the building design and construction. And similarly, for the compute power integrated into the car or an optimus bot...it is again the cooling and power requirements that need to be worked in the design. So maybe, from the persepective of Elon/Tesla, power just ends up being the metric that they focus on the most.

*Edited to add more thoughts.

In the traditional world of high-performance computing (HPC), the annual Top 500 supercomputer list often
includes power ratings for the installation:

June 2024 | TOP500

top500.org

The top installation by power is at 38.7 mW at Argonne National Laboratory. The power budgets are for unified systems which are not a dog's breakfast of installations in Texas that Elon is hinting at.

[Aside: One interesting setup is in an old church which houses Spain's largest supercomputer, listed as #8.
I just returned from a vacation there, but it was unfortunately closed for visitation.]

Barcelona church hosts supercomputer

It might look like something out of “Mission: Impossible,” but you don’t have to sneak inside like Ethan Hunt to catch a glimpse of this supercomputer in action.

www.lonelyplanet.com

Mind, the petaflop/exaflop counts are based upon 64-bit floating-point LINPACK benchmark for standardization
(using trad n**3 matrix multiplication math, not the Strassen method etc.) Although these national lab installations
are topping out at 1.5-2 exaflops, using 16-bit FP16 math they are more like 18-20 exaflops, which is closer
to the marketing flops of Tesla et al. (9 petaflops per training tile?).

See below for a nice article on how a better equivalence between FP64 measurements and the AI-oriented workloads may obtain, along with cost-performance tidbits relating to their installation contracts. Note that the AMD Instinct series (MI250 transitioning to MI300X GPUs) is well-represented along with Nvidia H100 accelerators in the Top 500. The first 2 FP64 exaflop machine will be Livermore's El Capitan using AMD; they are willing to spend 30-35 megawatts (and ~500M capex) on this, while they are not short on power due to their laser fusion efforts.

Lining Up The “El Capitan” Supercomputer Against The AI Upstarts

The question is no longer whether or not the “El Capitan” supercomputer that has been in the process of being installed at Lawrence Livermore National

www.nextplatform.com

About power efforts again, here is Oak Ridge (ORNL) chatting about their 30 megawatt installation using a new
power plant, while *all* of ORNL uses 150 mW for their entire campus.

HPE and AMD Powered Frontier Tops the Top500 Ushering in Exascale

Today we have the new ORNL Frontier supercomputer taking the top spot on the Top500 using HPE Cray EX with Slingshot and AMD CPUs and GPUs

www.servethehome.com

So, Elon humblebragging about 150-500 mW peak cooling must be Texas talk, but it's fairer to normalize
against actual computing power.

 

[MP3Mike]

So, Elon humblebragging about 150-500 mW peak cooling must be Texas talk, but it must be normalized
to actual computing power.

Elon isn't talking about just power for cooling, he said compute as well.

 

[insaneoctane]

For us old-timey, efficient market types, could you maybe explain what you're talking about here? And do you have any sort of prediction for TSLA stock price move tomorrow as a result?

Sure. It will close at $180. There's a lot of money to be made ensuring that TSLA closes at "maximum pain" on any given Friday. But tomorrow is "triple witching" Friday. Triple witching is the simultaneous expiration of stock options, stock index futures, and stock index options contracts, all on the same trading day. This happens four times a year, on the third Friday of March, June, September, and December. The expected expiration date for the three can increase trading volume and cause unusual price changes in the underlying assets. Let's see how close they stick the landing at the close to $180 tomorrow...

 

[SO16]

The media is sure effective in Australia. Would be great if Tesla finishes the year with strong sales.

 

[spacerfc]

Can someone explain to me why Elon talks compute in terms of Energy?

AI is creating huge demand for electricity, the power plants and the grid could be a limiting factor if they couldn't keep up:

How AI Is Fueling a Boom in Data Centers and Energy Demand

AI is causing the rapid construction and usage of data centers around the world—potentially imperiling climate goals.

time.com

Data center, AI load growth could threaten grid reliability: Conference Board

Although AI tools could eventually enable more efficient electricity generation, storage, transmission and use, near-term AI demand could slow the energy transition, the report says.

www.utilitydive.com

500MW is a modest sized power plant, he needs to have a plan for getting this power (and it needs to be continuous power, so if they go with solar/wind, it needs sizeable batteries).

 

[Snerruc]

AI is creating huge demand for electricity, the power plants and the grid could be a limiting factor if they couldn't keep up:

How AI Is Fueling a Boom in Data Centers and Energy Demand

AI is causing the rapid construction and usage of data centers around the world—potentially imperiling climate goals.

time.com

Data center, AI load growth could threaten grid reliability: Conference Board

Although AI tools could eventually enable more efficient electricity generation, storage, transmission and use, near-term AI demand could slow the energy transition, the report says.

www.utilitydive.com

500MW is a modest sized power plant, he needs to have a plan for getting this power (and it needs to be continuous power, so if they go with solar/wind, it needs sizeable batteries).

Which they already have in operation.

 

[Mengy]

The media is sure effective in Australia. Would be great if Tesla finishes the year with strong sales.

That was an interesting video. I find it hugely ironic that nearly none of them knew Tesla MY was the best selling car last year, but almost ALL of them knew Tesla sales had gone down this year. And a few parroted the battery lifespan concerns and other untrue negatives about EVs. The "Media" is doing a fantastic job with the FUD but not a very good job with positive news I see.

 

[FastEddieB]

Tomorrow is Friday. The MM and Hedge Funds will try to close at Max Pain because it costs them money if it's higher or lower.

MM=Major Market?

So, the prediction is market manipulation favors a close today at or near 180? Won’t take long to see if that prediction holds up.

 

[cliff harris]

MM means market makers. They are basically middle men who agree to always be prepared to buy any stock and sell any stock, so there is always liquidity in the market. This is why you can buy or sell any stock instantly, without waiting for a seller or buyer to also want to trade. Of course the market makers do very well on the 'spread' between the buy and sell price, which is how they get paid for providing liquidity.
It also means they have near perfect information, on all markets in which they participate. They see ALL the orders, so are in a very informed position from which to manipulate stuff.

 

[CHGolferJim]

MM=Major Market?

So, the prediction is market manipulation favors a close today at or near 180? Won’t take long to see if that prediction holds up.

FYI, Papafox’s daily thread includes a chart, at least once a week, that shows individual Friday closing prices vs. MaxPain over a period of months. They rarely diverge by more than $5 (MaxPain can change daily depending on recent call and put trading).

 

[Tony73]

You/mods should really move these posts/help/info to the Energy portion of this forum as and it's where it is more relevant. Someone else maybe in your situation or has done your exact same thing and can discuss their solar roof/installer. In terms of investment for TSLA, their home consumer energy market has never been that well supported/strong at all, and this is someone who really wanted to buy the Tesla PW. It's also not technically advanced IMO. Outside of folks who had contracts for Tesla to install solar roof at those old prices, you're at their whim of install areas.

From what I've read, Tesla won't even install on a metal roof so you are at the mercy of someone else already. There are also other solar tile/roof makers actually you can research. Tesla isn't the only game in town, even more so for consumer energy.

Tesla should exit the solar business. All it’s doing is causing reputational damage and wasting money. If the future really is Optimus and robotaxi, wasting resources on solar makes even less sense. It seems Musk would rather keep up the charade than admit the Solarcity acquisition is a bust

 

[Tony73]

I got some smiley faces a few weeks back when I said that maybe Nvidia should be worried about us instead of us worrying about Nvidia. Given huge margins at Nvidia, isn't it possible that we can undercut them when it comes to training ... especially if we have better software? (I freely admit my ignorance in this area and ask the question again in the hopes it will stir some debate.)

https://twitter.com/x/status/1803849373018644880

Edit:
Then HW5, which has been renamed to AI5, in the second half of next year. The Tesla AI5 computer has ~10X the capability of HW4 computer and Tesla makes the whole software stack. Elon

Nvidia makes its living selling chips. Its huge margins come from its HW business. It’s just starting to dip its toe into cloud services.
I seriously doubt Tesla will ever get into selling chips.

 

[SOULPEDL]

Sure. It will close at $180. There's a lot of money to be made ensuring that TSLA closes at "maximum pain" on any given Friday. But tomorrow is "triple witching" Friday. Triple witching is the simultaneous expiration of stock options, stock index futures, and stock index options contracts, all on the same trading day. This happens four times a year, on the third Friday of March, June, September, and December. The expected expiration date for the three can increase trading volume and cause unusual price changes in the underlying assets. Let's see how close they stick the landing at the close to $180 tomorrow...

Sticking pretty close to Max so far but it's up from 175 earlier in the week.
They are trying to keep it down so hard, my thinking is it will unwind early next week, so today might be a good buy still.

 

[growler23]

Seems like they have enough thermal processes at that plant to make use of that heat. More heat is the last thing Austin needs.

From what I've read, low level waste heat (which it seems like that heat is since the temperatures involved will be far lower than hundreds of degrees C) is pretty hard to use.
Given the mission and their obvious previous attention paid to it (clever Gigafactory energy efficient design, massive solar investment, water capture and reuse at Giga Texas) I have high faith their engineers have looked into whether/how to reuse it.

Spoiler: Possibly useful engineering background

I couldn't find the exact link, but my impression about the difficulty of using low level waste heat comes from Michael Barnard over at the CleanTechnica site. His stuff is worth reading - he does back-of-napkin analysis of a lot of clean energy startups, many of which he debunks immediately using that analysis. In some cases, several years later, after non-engineering-savvy investors have lost millions, he is proven right. A classic debunking he made early was of that startup (anyone remember the name?) that was going to store energy by raising a lowering concrete blocks using cranes. Which IIRC tried to get a second round of funding by replacing the concrete blocks with retired compressed wind turbine blades. Not sure whether Trevor Milton was involved in that one, since at least the physics could work, if perhaps not the economics.

 

[FastEddieB]

FYI, Papafox’s daily thread includes a chart, at least once a week, that shows individual Friday closing prices vs. MaxPain over a period of months. They rarely diverge by more than $5…

Couldn’t a random walk also result in a close +/- $5?

 

[cliff harris]

From what I've read, low level waste heat (which it seems like that heat is since the temperatures involved will be far lower than hundreds of degrees C) is pretty hard to use.
Given the mission and their obvious previous attention paid to it (clever Gigafactory energy efficient design, massive solar investment, water capture and reuse at Giga Texas) I have high faith their engineers have looked into whether/how to reuse it.

Spoiler: Possibly useful engineering background

I couldn't find the exact link, but my impression about the difficulty of using low level waste heat comes from Michael Barnard over at the CleanTechnica site. His stuff is worth reading - he does back-of-napkin analysis of a lot of clean energy startups, many of which he debunks immediately using that analysis. In some cases, several years later, after non-engineering-savvy investors have lost millions, he is proven right. A classic debunking he made early was of that startup (anyone remember the name?) that was going to store energy by raising a lowering concrete blocks using cranes. Which IIRC tried to get a second round of funding by replacing the concrete blocks with retired compressed wind turbine blades. Not sure whether Trevor Milton was involved in that one, since at least the physics could work, if perhaps not the economics.

I believe that was gravitricity, although I did not know that the economics were shaky. It might be another one. Gravitricity was re-using old mine shafts as I recall.

 

[Snerruc]

There is a hack that has shut down the dealer network almost completely. Also, several states use the company for registration. This means Tesla is probably done for the month in those states. Knock a few thousand off Tesla’s sales this quarter.

 

[Tony73]

I don't think Nvidia got to a $3.3T valuation by just making chips. Isn't the value of Nvidia in what they can do with their chips? Isn't that where we'll be bumping heads with them? If, if ... we get other automakers to sign up for FSD won't we be the leader in the licensing of full-stack real-world AI? To be followed by the licensing of multiple different bot embodiments and applications? Isn't that a bigger TAM than just making chips? Isn't that where the "durable competitive advantage" lies in the long term? I don't want to build 10% of the bots out there - I want to own the software and hardware they run on.

NVDA literally got to a $3.3T valuation by selling generalized AI chips supported by their proprietary software ecosystem.

https://www.visualcapitalist.com/nvidia-revenue-by-product-line/#:~:text=powered%20by%20Syntax.-,Advertisement,fourth%20quarter%20of%20last%20year.

Nvidia makes more gross profit from 1 H100 chip than Tesla makes from a model Y.