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

[StealthP3D]

Except he was already looking at selling shares to cover the bill.

If you just sell shares, you lose shares. 100% for sure.

If you instead sell a covered call you not only might not lose shares (but still get money), if you do lose them you get more money for them (both the premium, and the higher strike price, as compared to simply selling the shares outright.

In practice, it doesn't work like that. Because you have to sell enough covered calls to cover your tax bill with just the premiums in case the share price doesn't rise past the strike price. But if it does, then you have sold far more shares than you needed to, thus taking the problem you were trying to solve with this strategy (not wanting to sell any shares) and multiplied it many-fold. And the higher the strike you choose (in order to minimize the chances of the shares being called) they more shares you lose if they are called (because the premium is smaller and you need to sell more of them to pay your tax bill).

All you have done is traded the CHANCE to not have to sell any shares for the CHANCE that you might have to sell many times more shares than had you simply sold the amount necessary to cover your taxes.

There are no free lunches in this game.

 

[Singuy]

In practice, it doesn't work like that. Because you have to sell enough covered calls to cover your tax bill with just the premiums in case the share price doesn't rise past the strike price. But if it does, then you have sold far more shares than you needed to, thus taking the problem you were trying to solve with this strategy (not wanting to sell any shares) and multiplied it many-fold. And the higher the strike you choose (in order to minimize the chances of the shares being called) they more shares you lose if they are called (because the premium is smaller and you need to sell more of them to pay your tax bill).

There are no free lunches in this game.

Wait your tax bill is off the premium or your entire covered call value?

 

[StealthP3D]

Wait your tax bill is off the premium or your entire covered call value?

Since you would be the one selling the calls, they would be one and the same.

 

[wipster]

Did you go rate shopping, and if yes, who was the best? I attempted to take a collateral loan against shares and I thought it was going to be funded but they came back and said "no" because I was not diversified. Previously, they said they would, but only 30% of the value if not diversified.

No, when I got a good rate from my broker eTrade, I just followed through there. Heck, it was just too easy. Have to do it through their website, not the app, but I was approved almost instantly. Wish I could remember the name of the member who recommended this path to me, but my memory is like a steel trap with the door left open.

Happy hunting!

 

[MC3OZ]

Rough guess

Cost of Samsungs 12inch 14nm Wafer is $2000(it was 2500 in 2017 so I'm assuming it cost less today)
You can make ~200 FSD chips per wafer(assuming 85% yield)
Cost per FSD Chip is 10 dollars.
PCB is about 50 bucks

Total cost to Tesla is between 60-100 bucks per FSD computer.

Nvidia charges 2500/Xavier. That's their L2 stuff with 30TOPs. Their L5 stuff doesn't have a public price, but it's suppose to be 320TOPs. Knowing Nvidia, they will charge at least 5k to 10k for something like this.

So that's the Tesla FSD advantage. 100 bucks at most compared to who the heck knows what Nvidia charges. Plus they probably license out their software for additional milking.

You need to take charge of your own destiny or else you're just another automaker.

We can also question whether a faster processor makes a better FSD soluion.

It is important that the processor is adequate, Telsa intended HW3 to be adequate. Only staff closely involved with the project know how much of the processor is currently being used.

Tesla also did an buy out / hire of a small team who were experts in tuning NNs for lean use of resources.

When we come to HW4 it is supposed to be 3X the speed of HW3 but my hunch is lower cost and lower energy use are equally important targets.

When Tesla is shipping over 1 million cars per year these cost savings add up.
There is a synergy between scale and cost. For low volumes a off the shelf generic solution might be best.

For fitting FSD to a $25K Robo-taxi cost is all important.

The $25K Robo-taxi will be hard to do even with HW4, castings, structural battery, 4680s etc.

How a competitor could manage to build a $25K Robo-taxi anytime in the next 5-7 years is hard to imagine, especially if they use lidar.

In the meantime using 1 million plus FSD chips each year provides strong incentive for Tesla to keep improving the product, especially if those improvements lower costs.

 

[Buckminster]

https://twitter.com/elonmusk/status/1348043839022370817

Maybe a Tesla AI day later this year

New thread:
Tesla AI day

 

[aubreymcfato]

I also love this idea of Model Q. Honestly for a small car Q sounds.... cute.

Can we please remember that Qanon believes in this mysterious Q guy and we really do not want for Tesla to be associated with this crowd? It's the same crowd that stormed the capital few days ago.

 

[LiveLong&Profit]

If they didn't want us tough, they shouldn't have put us through so much the last 5+years.

Sold a covered call 12/28 @ $900 with greater than 50% conviction it would slide to Jan 2022 expiration worthless. 9 trading days in.....it's hanging by a thread. Lol!

This company is taking over the world, and the world will be much better off for it! Even when I lose, I win.

+100

 

[Stormy24]

Rough guess

Cost of Samsungs 12inch 14nm Wafer is $2000(it was 2500 in 2017 so I'm assuming it cost less today)
You can make ~200 FSD chips per wafer(assuming 85% yield)
Cost per FSD Chip is 10 dollars.
PCB is about 50 bucks

Total cost to Tesla is between 60-100 bucks per FSD computer.

Nvidia charges 2500/Xavier. That's their L2 stuff with 30TOPs. Their L5 stuff doesn't have a public price, but it's suppose to be 320TOPs. Knowing Nvidia, they will charge at least 5k to 10k for something like this.

So that's the Tesla FSD advantage. 100 bucks at most compared to who the heck knows what Nvidia charges. Plus they probably license out their software for additional milking.

You need to take charge of your own destiny or else you're just another automaker.

FSD chip has a copy of itself (2 x FSD chips) in case of a problem. So do they have two NVIDIA chips on there pcb?

 

[winfield100]

Schwab calls it a Pledged Asset Line (PAL). I don't know anything about it, but I'm also interested in learning more about this option.

@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"

a snip
".....Maintain a buffer of additional eligible collateral in the Pledged Account.
At the time that you establish the PAL, pledge eligible collateral in excess of the amount required to support the line of credit.
At all times thereafter, in order to prevent a Demand, maintain eligible collateral in excess of the amount required to support the greater of the minimum loan value of collateral and outstanding loans.
Maintain a diversified portfolio in the Pledged Account.
Maintain a diversified portfolio of eligible collateral in the Pledged Account as collateral support for PAL obligations.
Immediately address a Demand.
Consider keeping a readily available liquid cash reserve on hand in order to take immediate action and address a Demand......."

AND
"......Proceeds must be used for lawful purposes and may not be used to purchase securities or pay down margin loans. Additionally, PAL proceeds may not be deposited into any brokerage account......"

 

[bkp_duke]

I don't believe it's wise to underestimate Nvidia here. They are running machine learning for Nvidia Drive on a Top 10 supercomputer in their HQ in Santa Clara. Nvidia has a long, long reputation of being underestimated by competitors which no longer exist today. Tesla is likely to get to FSD first, and Nvidia will likely be second, and Nvidia is going to be the one selling their platform to all the non-Tesla companies. Tesla will be the Apple of FSD, but Nvidia will be the Android.

Tesla moved away from NVIDIA to their own solution.

I'll let that sink in for a minute.

 

[BOBTHEJOCKEY]

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?

 

[winfield100]

I believe the real value in Powerpacks at Superchargers is not for rate arbitrage but rather for peak shaving. Utility demand charges can drive up costs by several times.

@scaesare
perhaps also as nodes in local VPP's as PV and battery combo's build out.
Have you gotten your roof covered with PV and a battery in the garage yet?
we got a 1.13kw system in 1999 and a 11,655 system 19 years later @1/4 the cost and 10x the power.
battery is awaiting me convincing the boss we need one,
so the Cybertruk will be my deft, sneaky way of getting my 1/2 mobile powerpack
Floriduh does still have hurricanes and I expect VPP nodes will start popping up

 

[Maarten]

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?

Why are you shouting?

 

[winfield100]

1,500% here.
That's 167% CAGR.
I will celebrate by HODLing.... eh?.... everything.

@Artful Dodger
(1,390.06%)
2-Year Change
(and I would have at _least_ 5x as many shares if I had not been "smart" and traded instead of just HODL. I could have had almost a block of shares, but _no_)

 

[BOBTHEJOCKEY]

I highly doubt they would go after a company that has single digit market share

CORRECT but if Tesla gets ROBOTAXIS first it will have 100% monopoly on one game changing industry. This alone would correctly propel Tesla above $1000 and into the range of worlds largest market cap inviting in the TRUSTBUSTERS soon after.

 

[Singuy]

FSD chip has a copy of itself (2 x FSD chips) in case of a problem. So do they have two NVIDIA chips on there pcb?

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.

 

[Fobble]

Apple never had the best hardware spec on their phones comparing to the rest of the industry. But because they operate such a closely controlled hardware-software environment, they are able to optimize their phones in levels other phonemakers can only dream. Same idea with TSLA as they keep a tight control on their hardware and software design. And it's a luxury to have when you are able to attract the best talents.

Actually Apple has the worst hardware of all cell phones, more or less including the cheap ones. I recalled its recent models still use dual cores with 2-3MB memory, while even the cheapo Chinese phones had moved to Octa cores and 3-4GB memory way before that. However, on application and benchmark testings, Apple phones perform the best. Customization and optimization are real. Cheap hardware and high performance that is what boosts Apple profit margin. This is also Tesla's business strategy.

 

[Singuy]

Tesla moved away from NVIDIA to their own solution.

I'll let that sink in for a minute.

We can't say it's because Nvidia sucks. But we can def say he applied first principles and figured that a Tesla designed pcb at most cost a hundred bucks which is scalable to 20 million cars when Nvidia being the monopoly in town charging ransom money.

 

[bkp_duke]

We can't say it's because Nvidia sucks. But we can def say he applied first principles and figured that a Tesla designed pcb at most cost a hundred bucks which is scalable to 20 million cars when Nvidia being the monopoly in town charging ransom money.

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.