As my GOAT 🐐 once said, find the side who's coping the most, and make sure you're not on that side.
Honestly, this is becoming one of my favorite posts in here, and not just because of what Elon's saying. It's more so that if you trace the comments, the quotes and the threads on a post like this, you just find a lot of cope. Elon is frankly stating the difficulties of getting credit / capital, and who's going to pay for it. You basically buy the hardware up front, and the revenue only starts coming in once it's installed; and because you're borrowing money, each delay is hurting you as interest piles up and you're not bringing in any cash. Especially with the size of money you're dealing with, financing is now becoming the "bottleneck."
Now, for the uninitiated, what they see is only one side of the coin. They're saying, "Oh, demand is real," and if people can't build these data centers and there are delays, it's actually bullish, because that means the current players who have the compute up can charge a lot more, and therefore everybody's happy. No. That's not how you want to see this dynamic. This is late cycle, and it's best to accept it for what it is.
What I mean by this is Elon's basically raising the question of who's gonna pay for this and when; and not only that, can you get the capital, and on what terms? And even if you do get the capital, can you build out fast enough to take advantage of elevated spot GPU pricing? Also, what you're forgetting is the builders of these things are quite levered themselves, so all of this, at the end of the day, depends on raising money. This even goes to the labs: even $ANT, which of course raised a lot of money with the IPO, but we know right now that their credit is not seen as investment grade. Of course, that's subject to change.
So no, it's not actually bullish. A customer who can't get the money is the same as a customer who doesn't want the chips; either way, the order doesn't get placed. And the delays do not prevent a glut; they create one, actually. Supply is already built for the fast schedule, so when the buyers can't actually turn on what they bought, they stop ordering the next batch and it all piles up.
I know this sounds a little unintuitive, and you might have to think about it for a few moments.
This is why, to me, it's kind of tragic in a sense that a lot of the gains of the cycle have come from semi. I'm talking in terms of equity; if you're new here, I'm only talking about equity here.
I think we bounce in the next week; by how much, I can't say. And I want to say stocks don't go in a straight line, but if you look at this month's tape, and honestly since October, you just gotta kind of laugh. I guess stocks do go down in a straight line. But we've seen these names have very strong bounces. I want to say 2 weeks ago on that Wednesday, we saw AI names up 15%, 18%, just to pare it all back and then some. And for $NBIS, we've actually seen it multiple times be up like 5% pre-market, get sold off all day, go red, and go down even some more.
So it's actually fairly difficult to execute on this, because you don't want to get completely washed out, so you kind of need duration. And because this is so unintuitive on some level, you can easily cope and delude yourself that, hey, demand's so strong, etc. You're gonna be talking past yourself. This doesn't have to do with demand right now per se; this is a credit issue, and you have to look at it that way, in my opinion. If you just focus too much on demand, pie in the sky ideas, singularity, etc well, I hope you really believe in those things, because you're about to lose a lot of money. But that's okay, men have died for their beliefs in the past so what's losing some money?
I personally like having my head on my shoulder so I'm going to try my best to keep it on.
Getting this amount of compute turned on, cooled and working reliably is incredibly hard.
The cash shortage is also going to be nuts next year. Even cumulative industry delays of a few months between paying for compute and getting it reliable enough to generate revenue will mean hundreds of billions of incremental dollars just for working capital. Where will that come from?
@schemington Why ive been trying to bid google. I think im too early.
I wouldn't say in banging on the table (they could totally double down and not cut capex).
But i think it's good long, especially to balance out shorts.
As my GOAT 🐐 once said, find the side who's coping the most, and make sure you're not on that side.
Honestly, this is becoming one of my favorite posts in here, and not just because of what Elon's saying. It's more so that if you trace the comments, the quotes and the threads on a post like this, you just find a lot of cope. Elon is frankly stating the difficulties of getting credit / capital, and who's going to pay for it. You basically buy the hardware up front, and the revenue only starts coming in once it's installed; and because you're borrowing money, each delay is hurting you as interest piles up and you're not bringing in any cash. Especially with the size of money you're dealing with, financing is now becoming the "bottleneck."
Now, for the uninitiated, what they see is only one side of the coin. They're saying, "Oh, demand is real," and if people can't build these data centers and there are delays, it's actually bullish, because that means the current players who have the compute up can charge a lot more, and therefore everybody's happy. No. That's not how you want to see this dynamic. This is late cycle, and it's best to accept it for what it is.
What I mean by this is Elon's basically raising the question of who's gonna pay for this and when; and not only that, can you get the capital, and on what terms? And even if you do get the capital, can you build out fast enough to take advantage of elevated spot GPU pricing? Also, what you're forgetting is the builders of these things are quite levered themselves, so all of this, at the end of the day, depends on raising money. This even goes to the labs: even $ANT, which of course raised a lot of money with the IPO, but we know right now that their credit is not seen as investment grade. Of course, that's subject to change.
So no, it's not actually bullish. A customer who can't get the money is the same as a customer who doesn't want the chips; either way, the order doesn't get placed. And the delays do not prevent a glut; they create one, actually. Supply is already built for the fast schedule, so when the buyers can't actually turn on what they bought, they stop ordering the next batch and it all piles up.
I know this sounds a little unintuitive, and you might have to think about it for a few moments.
This is why, to me, it's kind of tragic in a sense that a lot of the gains of the cycle have come from semi. I'm talking in terms of equity; if you're new here, I'm only talking about equity here.
I think we bounce in the next week; by how much, I can't say. And I want to say stocks don't go in a straight line, but if you look at this month's tape, and honestly since October, you just gotta kind of laugh. I guess stocks do go down in a straight line. But we've seen these names have very strong bounces. I want to say 2 weeks ago on that Wednesday, we saw AI names up 15%, 18%, just to pare it all back and then some. And for $NBIS, we've actually seen it multiple times be up like 5% pre-market, get sold off all day, go red, and go down even some more.
So it's actually fairly difficult to execute on this, because you don't want to get completely washed out, so you kind of need duration. And because this is so unintuitive on some level, you can easily cope and delude yourself that, hey, demand's so strong, etc. You're gonna be talking past yourself. This doesn't have to do with demand right now per se; this is a credit issue, and you have to look at it that way, in my opinion. If you just focus too much on demand, pie in the sky ideas, singularity, etc well, I hope you really believe in those things, because you're about to lose a lot of money. But that's okay, men have died for their beliefs in the past so what's losing some money?
I personally like having my head on my shoulder so I'm going to try my best to keep it on.
Getting this amount of compute turned on, cooled and working reliably is incredibly hard.
The cash shortage is also going to be nuts next year. Even cumulative industry delays of a few months between paying for compute and getting it reliable enough to generate revenue will mean hundreds of billions of incremental dollars just for working capital. Where will that come from?
@J5jayfive I was battling with those thoughts for like 6 weeks i don't have great answers for you to be honest. I basically gave in and saw it as a credit issue.
I remain very enthusiastic about the technology, but I'm very cautious with captial right now.
@J5jayfive Yeah pretty much. I still have some sndk i bought last year, but im no where close to being OW AI names.
BUT im super excited for the technology, going to add the gemini $100 plan! I'm geeked right now i hope the technology keeps improving.
So bullish AI , bearish semis
I'm thinking out loud here. I'm pretty certain we get a bounce early next week, but maybe pop and fade. I just don't see how names like $CBRS and $ARM don't take a 50% haircut in the year of our Lord 2027.
Even with $SNDK, even though I'm a holder, I honestly can also see a haircut next year. It's really tough right now because we're seeing them fall from so high, but if there's a glut, you have to hunker bite the bullet and just be ready for that downside. And if you can't short, then don't go long.
Reason for this post is realizing $MU
stock bottomed 2 years before the earnings.
80% drawdown from 1995 to 96, but the worse print wasn't until 1998
So you were probably coping using demand and PE, etc all the way down.
And yes that's what i think is happening now.
I'm thinking out loud here. I'm pretty certain we get a bounce early next week, but maybe pop and fade. I just don't see how names like $CBRS and $ARM don't take a 50% haircut in the year of our Lord 2027.
Even with $SNDK, even though I'm a holder, I honestly can also see a haircut next year. It's really tough right now because we're seeing them fall from so high, but if there's a glut, you have to hunker bite the bullet and just be ready for that downside. And if you can't short, then don't go long.