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

[Singuy]

I didn't say NVIDIA sucks. But it is clear that Tesla determined that even with having to hire their own chip-development team, they thought they could do substantially better. And chip design, especially custom circuits, is one of the most difficult tasks there is.

His chip development team was some engineer ninjas because Nvidias r&d budget is 2.8 billion and Tesla's R&ds budget grew only 0.7 billion from 2015 to 2018 when the chip, model 3, solar tiles and future cells were being developed. And it'll take 2022 before Nvidia can surpass Tesla in performance per watt on a much superior node..like wtf kind of money management is that? Only this type of money management can result in a rocket company that doesn't go bankrupt instantly.

GMs R&D is almost at 7 billion a year. Like wtf are they working on because certainly it's not to increase margins. So where is the value generation to investors.

One metric I use to gauge my investment is on performance/r&d unit. This tells you the efficiency and competency of the company. This is why I went heavy on AMD which has 1/20th the R&D money of Intel/Nvidia combined and yet has products that are comparable and now in the processor space, better than Intel.

 

[ZachF]

They are now on FSD beta NINE. We all know it is awesome bla bla bla. Why are they waiting so long for us peasants to get it and what will be the FSD PRICE INCREASE when we actually do get it? I believe this is the ONLY THING that will really move this BUBBLE STOCK. upwards. Do you agree that the longer they wait the better chance that cnn msnbc fox will make BREAKING NEWS about TESLA FSD instead of TRUMPANZEE stuff. I actually agree that the longer we wait the more the conversation will be positive about ROBOTAXIS. Does Elon want to wait until the only Trumpanzee argument is that the devil and QANON is actually driving Tesla cars and NOT science?

 

[ZachF]

I didn't say NVIDIA sucks. But it is clear that Tesla determined that even with having to hire their own chip-development team, they thought they could do substantially better. And chip design, especially custom circuits, is one of the most difficult tasks there is.

Tesla/SpaceX/Elon (correctly) likes to keep as much of the value add under their own roof as possible, because first principles dictates a firm is truly only as valuable as the value it can create.

This contrasts sharply with modern MBA thinking where they think they can outsource everything, throw their badge on it, then magically mark it up.

 

[bkp_duke]

Tesla/SpaceX/Elon (correctly) likes to keep as much of the value add under their own roof as possible, because first principles dictates a firm is truly only as valuable as the value it can create.

This contrasts sharply with modern MBA thinking where they think they can outsource everything, throw their badge on it, then magically mark it up.

I am long aware of this. But they don't re-invent EVERYTHING. I.e. the cars still use Brembo brakes, 3rd party tires, etc. They have to see enough value in doing it themselves, and that must have been the case with chip design.

 

[Mike Smith]

If you think the average person understands TSLA, check out the comments in this non-EV subreddit on a thread about Tesla's market cap: https://www.reddit.com/r/dataisbeau...c_tesla_is_now_bigger_in_market_cap_than_the/

I read through about 50 comments and 100% of those people believe Tesla is a great company but wildly overvalued. This is exactly what people thought in 2019, and exactly what we want to see. I thought maybe people were starting to understand TSLA, but this gives me extreme confidence that we're still only in the middle of this AMZN-like growth phase in TSLA.

 

[Maarten]

Damn that's funny. I am a retired race horse jockey who posts a lot on Facebook under my real name Robert D'Amours. Horse racing people, who are my long time friends and tax clients, are quite smart but very uneducated and posses Trumpanzee intellect. Yes I do shout at them just like you do your Trump supporting neighbors. .

I live in The Netherlands. Never shout at my very friendly neighbors.

 

[BOBTHEJOCKEY]

Damn that's funny. I am a retired race horse jockey who posts a lot on Facebook under my real name Robert D'Amours. Horse racing people, who are my long time friends and tax clients, are quite smart but very uneducated and posses Trumpanzee intellect. Yes I do shout at them just like you do your Trump supporting neighbors. .

I read through about 50 comments and 100% of those people believe Tesla is a great company but wildly overvalued. This is exactly what people thought in 2019, and exactly what we want to see. I thought maybe people were starting to understand TSLA, but this gives me extreme confidence that we're still only in the middle of this AMZN-like growth phase in TSLA.

Yes but the game changer, propelling the stock above $1000, in my view is the Robotaxi and that's why Elon is waiting so long for us peasants to get the Beta zillionth download.

 

[Singuy]

I read through about 50 comments and 100% of those people believe Tesla is a great company but wildly overvalued. This is exactly what people thought in 2019, and exactly what we want to see. I thought maybe people were starting to understand TSLA, but this gives me extreme confidence that we're still only in the middle of this AMZN-like growth phase in TSLA.

Yup, I don't think people look into the numbers as closely as we do. Tesla will continue to surprise while other companies continue to shrink. Tesla is almost redefining the future of capex and r&d expenditure. It was blatantly obvious that Elon has already did this with spacex. Comparing r&d and capex budgets to other space related companies is almost laughable, especially to governments.

True story, I am very anxious to see the tron coaster get built at disney world. It broke ground Jan of 2019. It is now delayed to 2022. Tesla will have two giga factories built before I can ride this coaster..like wtf is that?

 

[Silent Ludicrosy]

Somebody on Twitter discovered their model Y has a single piece rear casting. The delivery date was Dec 23rd. https://twitter.com/tonytesla4life/status/1348099171899777024?s=21

 

[MikeQ]

I didn't say NVIDIA sucks. But it is clear that Tesla determined that even with having to hire their own chip-development team, they thought they could do substantially better. And chip design, especially custom circuits, is one of the most difficult tasks there is.

A custom tailored hardware solution for doing image processing and specific types of math will always be a better solution than a GPU. A GPU is a custom tailored hardware solution for doing 3D graphics. It just so happens you can write a shader for a GPU that can do image recognition. That doesn't mean it's the best solution for doing image recognition. It's better/faster to use the GPU than it is the CPU but a custom chip that can do image recognition in silicon is the best solution.

 

[Mike Smith]

Apple, Hyundai to agree on electric car tie-up early this year: Korea IT News

'Citing industry sources, the latest report said the pair plan to build the cars at Kia Motors’ factory in Georgia, or jointly invest in a new factory in the United States, with a plan to produce 100,000 vehicles in 2024 at the proposed plant with an annual capacity of 400,000 vehicles. Kia Motors is an affiliate of Hyundai Motors.

The report said Hyundai and Apple plan to release a “beta version” of Apple cars next year.

Both Hyundai Motor and Apple did not have any immediate comment.'

 

[MikeQ]

And chip design, especially custom circuits, is one of the most difficult tasks there is.

And Tesla had probably the best person for the job. Jim Keller from AMD. He's why AMD is now stomping Intel in the CPU space.

 

[StealthP3D]

@sroh
https://www.schwab.com/public/file/P-952913
it looks to me like a nice way to lose up to and more than _everything_ , kinda like options only under a different name
It kinda screams to me, if you want leverage here's a nice way to go broke and note, there are several (many) places where it says more or less, "keep excess cash on hand"

Yeah, I'm shocked at the people that think it's a good idea to take out a loan to pay capital gains taxes! In 30 years of investing I've never even considered doing that. The capital gains taxes come out of the sale that generated the gains. If your trading is not profitable enough to support that, maybe you should have thought of that before you made the trades!

It's one thing to take an extremely concentrated position in a high-flyer with money you can afford to lose but leveraging a concentrated position with borrowed money is just plain lunacy.

Get a grip people!

 

[BETA1019]

For those of you who were invested in AAPL way back in the day...

There was an instance (perhaps even two instances) where Apple issued some kind of official interim notice that they were essentially going to blow out their own sales forecasts for the iPhone.

Does anyone know the name of this kind of notice? I'm trying to look it up and see the historical trends on it, and compare it against what may happen to Tesla.

 

[traxila]

Apple, Hyundai to agree on electric car tie-up early this year: Korea IT News

'Citing industry sources, the latest report said the pair plan to build the cars at Kia Motors’ factory in Georgia, or jointly invest in a new factory in the United States, with a plan to produce 100,000 vehicles in 2024 at the proposed plant with an annual capacity of 400,000 vehicles. Kia Motors is an affiliate of Hyundai Motors.

The report said Hyundai and Apple plan to release a “beta version” of Apple cars next year.

Both Hyundai Motor and Apple did not have any immediate comment.'

As an FYI, I believe Hyundai has been collaborating with Canoo (GOEV) to produce their electric cars. Have no idea what this means with Apple in the mix. Is Apple bringing new technology to the table to ditch Canoo or is Canoo a huge beneficiary here?

 

[StealthP3D]

Yes, their level 5 stuff has two of everything. But unlike Tesla that spends another 10 dollars in cost, usually you pay exponential with third party. For example, a gtx 3080 is 699, but a 3090 is 1499 for 10% better performance.

It didn't look like the cost comparison you provided earlier accounted for two of everything. Was it apples to oranges?

 

[TheTalkingMule]

I read through about 50 comments and 100% of those people believe Tesla is a great company but wildly overvalued. This is exactly what people thought in 2019, and exactly what we want to see. I thought maybe people were starting to understand TSLA, but this gives me extreme confidence that we're still only in the middle of this AMZN-like growth phase in TSLA.

Supports my theory it’s gonna take another 3 years for the wider market to recognize and appreciate the value of Tesla’s place in the energy world. This was half the basis for my covered call strategy at the end of the year and my belief I’ll be able to buy back in if forced to sell at $1200.

At some point in the next couple years, a semi-reasonable multiple(400x) will be required and will be based only on automotive. Then the real Energy money will start rolling in around late 2022 and it’ll take a while for folks to buy into that exponential growth.

Edit: Sorry for discussing valuation/investment stuff in the investment thread.

 

[Singuy]

It didn't look like the cost comparison you provided earlier accounted for two of everything. Was it apples to oranges?

I said the xavier chip for Nvidia is 2500, which has only one chip. I also said their l5 pegasus doesn't have a listing price but that board has two of everything. Two Xavier's and two gpus. Nvidia however didn't give a price.

As for the Tesla side, the soc is one 260mm die which will cost 10 bucks per chip, or 20 bucks for the entire computer, hence I put down 100 bucks a board just to be conservative.

 

[scaesare]

A custom tailored hardware solution for doing image processing and specific types of math will always be a better solution than a GPU. A GPU is a custom tailored hardware solution for doing 3D graphics. It just so happens you can write a shader for a GPU that can do image recognition. That doesn't mean it's the best solution for doing image recognition. It's better/faster to use the GPU than it is the CPU but a custom chip that can do image recognition in silicon is the best solution.

Indeed.

What's more, there are a number of additional factors that have to be accounted for in a specialized dedicated application like an FSD computer: form factor, thermal envelope, latency, bandwidth, etc...

Each comes with a tradeoff, not unlike with batteries. Tesla made some very specific design choices, such as: limited but optimized instruction sets, cache design, SRAM placement, memory interconnect/bandwidth, etc... that allowed them to optimize the hardware.

Additionally they deliberately moved some features like layer fusion control, the ability to map SRAM, etc... in to the software/compiler. This allowed them to simplify the silicon, and provides flexibility for development, at the cost of increased software complexity.

Ultimately a specific approach like this allows them to meet the design goals along several axes in a way that relying on a more general-purpose system from a 3rd party would likely not.

 

[StealthP3D]

One metric I use to gauge my investment is on performance/r&d unit. This tells you the efficiency and competency of the company. This is why I went heavy on AMD which has 1/20th the R&D money of Intel/Nvidia combined and yet has products that are comparable and now in the processor space, better than Intel.

Smart. AMD has achieved 50-bagger status in the last 5 years! I recently found an AMD share certificate for 50 shares in a little used drawer that I had completely forgotten about. It's dated the day before the last split in 2000 but that wouldn't be the purchase date because we had an on-line brokerage account by 1996 or so. So it must be some orphan shares that were hiding in some closed-out account back when we used a broker we had to call on the telephone that got mailed to us right before the 2:1 split. So it's really 100 shares. Or maybe they are the result of a shareholder lawsuit I've forgotten about. I can see why these shares were easy to forget about because they have been worth as little as $150 in total as recently as 2015. All of a sudden they are pushing $10K.

See what happens when you forget you own shares? LOL!