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.
The conclusions in the article don't make sense to me, especially when they mention HPC chips in the article itself - HPC is datacenter, not in-car.

Only 25 chips from a 12" wafer at 7nm? That's probably a die size of over 400mm^2, maybe over 1000mm^2 depending on the assumptions you make about how good the yields are. That's as big or bigger than the largest high end GPU dies.

If 25 dies per 300mm (12") wafer are true, these are more likely chips for Dojo NN learning acceleration than chips for FSD HW NN in-car acceleration.

For comparison, HW 3.0 FSD chip is 20x13mm or 260mm^2, on 14nm node. On 7nm node, to be that much larger, would be an obscene number of transistors. A direct shrink from 14nm to 7nm would be roughly half the die area (in reality, nothing is a perfect scaling, it would be a bit bigger). So to be on the order of 3x~7x the transistor count (taking into consideration a smaller node and larger die area) would be nonsensical for FSD in-car hardware.

I suspect in their rush to be first to publish Electrek has made some incorrect assumptions.

I think the key is SoW: System on Wafer. Rather than just the HW3 chip, this die will have both chips (4 NN cores), plus GPUs, Arm cores, flash and dynamic ram. Possibly also the video front ends.
Basically, they pull in all the high end ICs from the board to one wafer/ package. This reduces parts count, packaging costs, and board size while increasing theoretical bus speed and width. Downside is yeild impact due to die size (but that can be worked around with redundant circuits).
They can use the wafter as a substrate (nano PCB) and mount memory (or other) ICs to it with wirebond connections.
The Next Advanced Packages
 
Make sure you don't have old limit orders in place folks. ;)

We've run up this high (well, relatively speaking) this fast twice this year already and we all know what came next. Will this time be any different?
If we hit close to 2k I think I'll have to sell one of my remaining calls for that reason. With shares I don't mind "leaving money on the table" from not selling and trying to time the dips. Today will be a great macro day. Big box guys blew it out of the water.
 
A bit of a try by the MM's to create some panic and a drop? not unusual in pre-market...

Now here's something that I was told by @NicoV over in the Tax-Split non-US implications thread.

Our broker advised to consider repositioning some calls as some of the new strike-prices could be a bit illiquid. At the time I didn't think about it too much, but actually I've got October calls for $1380 and $1820, the rest are all nice multiples (4/9 $1500, Jun 22 $1250, Jun 22 $3500), but those first two calls will be adapted to strikes of $276 and $364 - the options chain is going to be a mess and indeed, who'd be looking to buy on those strikes?

So thinking to roll these to a "nicer" number post-spltterino. Just wanted to gives the heads-up to the rest of you.
Well, they'll have weird strike prices, but there will be higher open interest and lower prices, so they shouldn't be too illiquid. What happened with the options in the SCTY acquisition was that the positions could only be closed, not new ones opened, and this did affect the liquidity somewhat. I wonder if that will be done again.
 
  • Informative
Reactions: Tslynk67 and CarlS
We've run up this high (well, relatively speaking) this fast twice this year already and we all know what came next. Will this time be any different?

Wild card here is the S & P. One would think they would be completely avoiding this until at least September, but stranger things have happened.

Anyone have any insight on what the short percentage might be at this point?
 
  • Like
Reactions: Nocturnal
3000 today I’m not even joking maybe $10,000

I have limit sell orders starting at 2000 all the way up to 10k
I did the same.

Just in case my theory few posts back (that 4/5 of float value disappearing next week, shorts need to deliver shares unless helped by brokers), is right.

That theory got a bunch of disagreements here, but I've learned to stick to my guns in life and on this forum too, when I think I'm right. I don't know that I'm right in this case, but I think it's a possibility. And even if technically I'm right, I think brokerages may do actions to avert consequences I was pondering, or even that markets end up halting Tesla for few days.
 
It's extremely depressing that most of the posts are unnecessary jokes, memes, gifs, youtube links to music, pictures of our personal wealth and non-Tesla-related stuff.
So euphoria is depressing? I've learned to just skip all the posts when I'm behind, I save a lot of time and keep making money.
 
Damn it...went away for a three day camping trip...came back to find a bazillion new post's.

Up side is I am now a lot richer.

Double upside is I will be even more richer (is the the correct way to say I am rich)...after the end of this week,

On the subject of retirement...well I am now officially retired and I am deep in the planning stages of what to do after I get tired of invading Island's.

Might look into electrifying a P51 mustang.
A P-40 would be cooler ;)
 
Last edited:
  • Disagree
Reactions: Unpilot
Due to some work pressures I'm about 70 pages behind!!! Still on Saturday :)

It's extremely depressing that most of the posts are unnecessary jokes, memes, gifs, youtube links to music, pictures of our personal wealth and non-Tesla-related stuff. If I were a moderator I would simply be deleting them with no warning or explanation... we ought to know that this thread is supposed to be about stock.

So CNBC reluctantly covered the stock just now. Instead of discussing what the company is doing, for example "they are expanding from 0.5 car million/yr to 3-4million/yr" it's "look how Elon Musk's wealth is increasing... he is now number four on the list." That is probably the only thing they can think of pin on him right now, pretty lame.

There’s several terms that could be used in this situation. I’ll start off with; stick in the mud.

We’ve analyzed to death (in fact I’m pretty sure we’ve all died at this point and have gone to simulation heaven, we just don’t know it) every detail of the company, the man, the technology, the SP, the shorts, basket weaving techniques, the batteries, the future second by second, the coded Twitter account, trains, planes and automobiles, the competition, the floor plans for Giga Nevada, Shanghai, and Berlin (next week Texas), the balance sheet, the unbalance sheet, the refresh, robo-everything, the criminals and bad actors, and on and on the list goes.

We’ve dissected all of it, ad nauseam. Like I literally have barfed.

It’s okay and refreshingly normal for people to take a break and enjoy life and the fruits of their labor. We’re winning on all fronts, so your ‘depression’ at the display of joy here simply solidifies my continued views that people suck and why nobody is invited to my island, ever.

Just to further make you depressed, because I’d never want to rain on your parade, my technical analysis is that the SP hits $2000 before the split because my mother called it two weeks ago out of the blue. She’s supposed to be dead according to many doctors and specialists, but she’s not. She’s quite literally a miracle. So I figure anything that comes out of her mouth is the word of The Universe.
 
There’s several terms that could be used in this situation. I’ll start off with; stick in the mud.

We’ve analyzed to death (in fact I’m pretty sure we’ve all died at this point and have gone to simulation heaven, we just don’t know it) every detail of the company, the man, the technology, the SP, the shorts, basket weaving techniques, the batteries, the future second by second, the coded Twitter account, trains, planes and automobiles, the competition, the floor plans for Giga Nevada, Shanghai, and Berlin (next week Texas), the balance sheet, the unbalance sheet, the refresh, robo-everything, the criminals and bad actors, and on and on the list goes.

We’ve dissected all of it, ad nauseam. Like I literally have barfed.

It’s okay and refreshingly normal for people to take a break and enjoy life and the fruits of their labor. We’re winning on all fronts, so your ‘depression’ at the display of joy here simply solidifies my continued views that people suck and why nobody is invited to my island, ever.

Just to further make you depressed, because I’d never want to rain on your parade, my technical analysis is that the SP hits $2000 before the split because my mother called it two weeks ago out of the blue. She’s supposed to be dead according to many doctors and specialists, but she’s not. She’s quite literally a miracle. So I figure anything that comes out of her mouth is the word of The Universe.
You know, maybe go easy on the hairballs.
 
The conclusions in the article don't make sense to me, especially when they mention HPC chips in the article itself - HPC is datacenter, not in-car.

Only 25 chips from a 12" wafer at 7nm? That's probably a die size of over 400mm^2, maybe over 1000mm^2 depending on the assumptions you make about how good the yields are. That's as big or bigger than the largest high end GPU dies.

If 25 dies per 300mm (12") wafer are true, these are more likely chips for Dojo NN learning acceleration than chips for FSD HW NN in-car acceleration.

For comparison, HW 3.0 FSD chip is 20x13mm or 260mm^2, on 14nm node. On 7nm node, to be that much larger, would be an obscene number of transistors. A direct shrink from 14nm to 7nm would be roughly half the die area (in reality, nothing is a perfect scaling, it would be a bit bigger). So to be on the order of 3x~7x the transistor count (taking into consideration a smaller node and larger die area) would be nonsensical for FSD in-car hardware.

I suspect in their rush to be first to publish Electrek has made some incorrect assumptions.
I thought the current FSD version 3 chip runs on a motherboard. Is it possible this chipset could replace the v3 chip and the rest of the board? Seems like a way to reduce power consumption significantly and reduce other chip costs and fabrication steps.
I am hoping this is for dojo, which I think is needed to get tougher corner cases. Could they be building a data center chip for dojo and contracting for v4 hardware in this same contract? Dojo will benefit more then cars with better performance chips and would need to cache much more data. V4 for the car should mostly be about reducing power, dojo will be about faster processing and storing massive data label libraries.