CRWV 是一个做空标的，也是破产候选：550 亿美元债务，而汇丰估算其投资资本回报率为 8.9%，低于其借贷成本，债券收益率为 15%，现金流很可能不足以偿债。与 WeWork 类似，其不断贬值的芯片所背负的债务无法随规模同步扩张。

You know, we're short from 107 earlier this year. We actually doubled the short a couple of weeks ago. It's it's our biggest position. I think there's a very good chance Cory goes bankrupt.

And I mean, Cory's already imploding. All right. It's only a question of time before this before this the whole structure the wheels are coming off the financing architecture of this whole edifice and therefore I think it's going to collapse.

Another short that I wanted to talk about is Cororeweave. Now this goes into the heart of our discussion which is the AI debt um boom which is becoming less of a boom now. Coreweave is one of your shorts. Um I believe you're up on that short as well.

Now um Cororeweave has $55 billion of debt. According to I was just reading this report on investing.com. HSBC is estimating that Cororeweave earns an average of 8.9% return on invested capital which is below supposedly below their cost of borrowing.

That's one issue. The other issue is that um uh they have around $35 billion of debt um 55 in total.

So if any of these companies start to uh and by the way Open AAI is not so far not profitable. So if any of these companies maybe pull back on financing potentially Cororeweave is in a lot more trouble.

If you look at the numbers, um the amount of the debt and the interest expense associated with that debt relative to the market cap and revenues of the company, it's just it's it's extraordinary.

Um and the problem with Cororeweave, I mean, there's there's they don't scale at all. There's no sign that the unit economics actually work.

Coral Corowe is basically we works with GPUs. You know, we works went bankrupt because why? They signed these long-term leases and they ran out space on a short-term basis to folks who need the office space.

Okay. Corweave, you know, has these these these these enormous uh uh this enormous exposure. They own a pile of chips, depreciating chips. All right, the debt and the depreciation are permanent and and they try to rent them out.

Now, in the short run, when the getting's good, they can rent them out at high at high rates, but looking out on anything but the short term, like what are the sustainable long-term economics of renting those chips out?

And uh it's it's it's rather questionable at best.

And so, you see spend, spend, spend. They're investing more and more and more. Debt's going up and up and up. Interest rates, expense up and up and up. Depreciation up and up and up.

But what's the end demand for these chips? what's the useful life of these chips? That's been a big big point of controversy.

So, um, no, I I I think Cororeweave, uh, you know, thank you for mentioning it. You know, we're short from 107 earlier this year. We actually doubled the short a couple of weeks ago.

It's it's our biggest position. I think there's a very good chance Corweed goes bankrupt.

that basically what's happened as the stock prices plummeted, retail investors at Robin Hood came piling in at record amounts to buy coral. This is not how bottoms are made. Bottoms are made when you have people saying, "Get me out.

There's blood in the streets. I can't take it." Retail has been embracing this with both arms, buying it to an extent they've never bought before.

cuz you look at the balance sheet most of the folks looking looking at Corave not paying attention to the balance sheet they're just buying the narrative. I don't think pe I think most investors in Corweave they don't even know what they own they just bought a story.

the central issue is that um they don't have they probably don't have enough cash flow to service their debt.

going back to my point, 8.9% ROIC is return on invested capital is what HSBC is estimating that the company will have for the next year and a half to to three years. But, uh, some of their bonds already reached 15%.

Which illustrates a point. Their cost of capital is a lot higher than the actual return on invested capital. So, that's one problem.

The other problem is why is the bond market signaling that they're in trouble when the equity uh market is is kind of reflecting that, not so much. I mean the entire AI sector is is more or less still in tech from a from an equities valuation standpoint.

I mean Cororeweave is down um from May uh yeah from it's at down 56%. Um but uh when you say this is an entire AI sector problem. I'm not seeing the entire AI sector selling off even though bond yields and financing costs are are really rising.

So what what is a bond market getting that the equities market is not yet getting?

So, so David, as you know, they often say the uh the bond market's smarter than the than than the equity market. The equity markets are dreamers. The bond market just wants to the bond guys just want to make sure they get paid back.

All right. >> Yeah. >> Usually, you're better off following the bond guys, not the equity guys. All right. So, if the bonds are not money good and the CDS are telling you that, okay, if the bonds are not money good, then the equity is worthless.

We're not quite there yet, but we're we're we're on the road, well on the way. And so, yes, I picked on the NEO class because that's the tip of the spear.

And so the bond market's telling you, the bond market's telling you that the equity is close to being worthless. And I believe it's only question of time before the equity will be worthless. I expect Cory to go bankrupt.

If let's say cororeweave does go under, what's the um next um link of the chain so to speak that's going to fall? >> Well, I expect other neoclouds will go down in sympathy. I expect and it's already happening.

The semiconductor stocks are taking on water big time. I mean, think about you, David, you read you read more than I do, okay? You saw the story overnight or yesterday. Uh it was going around a business inside somebody had Amazon has like $200 billion worth of chips sitting somewhere in a warehouse collecting dust.

It's like you know all these chips are being produced but they're not being put into service for various reasons. They can't they can't build they can't get the the permitting.

They can't do this. They can't do that. So there's a lot of double and triple ordering going on. >> This is like fake demand. It's demand brought forward. All right. And I I So what's the consequence of all this?

The semiconductor stocks are already falling apart. There's going to be more where that came from. I think I think the whole thing is going to burn down. I I really do.

The other side effect is what could happen to the banks and according to JP Morgan if hypothetically corweave defaults on its bonds. Um not much apparently. JP Morgan team said banks have helped to structure what are called delayed draw down or draw yeah draw down term loans or DDTL core can draw from these so credit lines either based on timeline or a project milestone such as the completed construction of a data center.

>> Okay. Coming just coming back to um financing for for uh uh for AI stocks and then I want to talk about your long side now. Um Nvidia back firm scraps $5 billion Australian IPO.

Um you know it's interesting I was just looking this up earlier according to prediction markets there's still a 60 cow on cowi there's still an above 60% chance that anthropic would IPO before the end of the year or announce an IPO before the end of the year.

Uh not high but not low either. But I'm wondering, you know, Fer Fermis, uh a data center operator backed by Nvidia in Australia just shoved it $5 billion IPO on Friday. Again, could be firm specific, but does this signal anything to you,

Yeah. And so you're maintaining your shorts basically. You're not you're not looking at these draw downs 50% of core for example. Um the text the semiconductors are falling, chip stocks are falling.

You're not looking at this as a buying opportunity. That's the bottom As a m as a matter of fact, you know, a couple weeks ago, we our new in fairness to the Noble Update subscribers, any positions that are less than 2 months old, we don't put the names in cuz we've, you know, our substack has doubled in the last couple of months because we we've been hot as hell with our picks.

But, uh, truth be told, a couple weeks ago, two weeks ago,