Transcription
If AI monetization gets pushed out and if it's not a 2027, 2028 occurrence but 2030 or whenever, then Oracle will have fundamental financial problems. You have a much riskier construction of the demand than you did in the telecom and .com era. The unprofitable companies as a percent of the spend is higher in this cycle than it was in the telecom cycle. As long as those unprofitable companies can keep raising money to keep paying those bills, that's fine. But if we get a credit crunch or we get a 2001, 2002 pullback in sentiment, you're going to see a lot of that spending drop and drop pretty quickly. It's just added risk. The risk levels for all these companies are much higher than they were now. I think in a way that's worse than 1999, 2000.
Today's episode is brought to you by Tukrium. Learn more about how agricultural commodities can play a role in your portfolio by clicking the link in the description.
Today we've got a very special interview. I am joined by Jim Chenos of Chenos and Company, investor, short-seller. Jim, welcome to Monetary Matters. The first question I want to ask you is, many of the investors in AI and in data centers, in GPUs, what do you think are some of their widely held beliefs that they are convinced is true, that you are doubting whether it is true or not?
Thanks for having me. So our view on the data centers is is is somewhat bifurcated. We know the actual data center legacy data center business pretty well. We put a a big short on it in mid-2022. And one of the things that struck us about the the cloud data center business, the old legacy data centers, which are much smaller than the the new ones we're talking about, is just what a what a crummy business it was. It was a very low return on capital business, highly capital intensive. The data center companies capitalize a lot of things that we believe should be expensed. They're negative free cash flow. I could go on and on and on and and we're still short those companies.
And so as as the AI story has progressed, the latest hot area is data centers, including now a view that we're going to be putting data centers in space, which I sort of joked about a year ago. It it of course, in this market, jokes become true. But what we we really pointed out was that that hosting GPUs is not a great business. It's it's basically a commodity business, particularly as everybody's building them. And our view to clients has been the magic and the money is going to come from what the chips produce ultimately, not where they reside. And that that if you're going to make a bet on AI, you should make it on either the pure AI companies like OpenAI, XAI, Anthropic, or you should you should bet on the hyperscalers and more on that later. But but investing in that Bitcoin miners that are converting to data center companies or or Meta companies that are jumping in or just the the so-called Neoclouds, which are basically just landlords, to me seems problematic for for two basic reasons. Hosting hosting GPUs or CPUs is inherently a low margin, low return on capital business, number one. And number two, you have to make a bet if you're actually buying the GPUs yourself, you then have to make a bet on depreciable lives, which is one of the big controversies. And how long until you have to replace the GPUs because your tenants want something bigger and better. And we can argue about that. I I have my own views. We use five-year life with 20% residual value. It it could be six years, it could be three or four years. Time will tell.
So if we had to say there's three different kind of stacks of this AI, there's the the hyperscalers who are selling things that are produced in the data center. There are the companies that sell things to and make the data center. Nvidia would be the most obvious of of this. And then the third are the companies that own the data center. It sounds like a a lion share of your skepticism and perhaps bearishness is concentrated in that in that third sector. How much would would that be concentrated in let's say the S&P 500? Because a lot of them are the hyperscalers, which is the first category, as well as Nvidia, which is the second.
So when we do our model short portfolio for clients, we we advocate that it always be hedged and that's how we ran our business since our fund business since '96. So I always joke I'm sort of long those companies because we we own the indices and they just make up such a huge part of the S&P and the and the NASDAQ. And so in effect, you know, I'm advising people to be long Microsoft and Nvidia and and Meta and and and Google. So it's really it's really the as you point out, it's the companies sort of the second or third derivatives of the companies that are that are trying to basically either recast their business or build a new business around owning the data centers and operating the data centers. That's the area we know well and we think is not a particularly good business.
And where does Oracle come into this?
Oracle is kind of a hybrid. You know, they they're they're a hyperscaler and of course they have a legacy software business that's profitable, but they're if you look at the the five hyperscalers that are building out data centers, which are Microsoft, Meta, Google, Oracle, and Amazon, they are one of the two that don't earn their incremental cost of capital. And that includes the numbers that were released last night. So in their zeal to sort of catch up to the others, I mean, they're they're expanding their balance sheet really, really quickly but don't have the income and cash flow levels that someone say like Meta or Microsoft do. Meta and Microsoft, by the way, according to our work, are the only two hyperscalers that are right now anyway earning above their cost of capital on their incremental capital spending. And and now that can change obviously as AI output hopefully gets monetized in a better way down down the road, but but Microsoft and Meta can easily handle their investment, their investment load. Whereas Oracle and and interestingly Amazon are at the bottom of that that stack in terms of their inability to at least monetize yet their their massive investments in in AI equipment and hardware and and locations.
Tell me about Meta. I had it referred to as a hyperscaler. I because I thought that meant, you know, providing clouds to to other people. I thought Meta was just spending all this money to improve Instagram and all these other things, or are they they are selling the data to other people?
Well, they they're they're integrating AI and building out massive data centers, including I think they're doing the big one in Louisiana, if I recall, with the controversial accounting, the SPV accounting. So, they they too are are building out most likely for their own internal capacity, their data center capital plan. And again, they're monetizing things through in effect selling ads to Instagram and Facebook users. And Microsoft is doing it through their corporate customers. So, at least for the time being, that they're the two that are that are able to sort of finance this from their own cash flows. The others are going to need external financing.
And that is a remarkable statement, Jim, because for so long, these Magnificent Seven companies have just been so profitable and have grown so much with so little capital requirement. Now, these these they're just chewing up capital and their free cash flow is, you know, steadily going down quarter over quarter. Are you envisioning a future where some of these companies free cash flow might go negative, meaning basically that their operating cash flows are are lower than the amount of money that they need to to invest?
Well, that's already happening with Oracle. So, they've they've gone free cash flow negative. I think it will erode. And again, the idea is is that AI monetization from output will, you know, hit hit an inflection point in 2027 or 2028 and then and then profits will flow. We'll see. I mean, I I you know, I I don't know and I I don't think anybody else really knows. We'll just have to see. That's the bet. The bet is a big one. And for some of them, it's a bet your company bet, which they never had to do before. The underpinning of your question is correct. In that you basically had asset-like businesses based on intellectual property that just gushed cash flow and now you have basically, you know, the second industrial revolution, if you will, where where the next round of technology improvement is really capital intensive and that's a change.
And how are you assessing that return on invested capital? You you can you sort of pick a company, but let's say 10 years ago when Microsoft was investing in building data centers for its, you know, earlier cloud clients. I presume the return on invested capital was quite high. What does the return on invested capital in this new AI data data center world look like?
Most of them don't break out their actual pure data pure AI revenue and costs and you can only look at the aggregate and and just make some assumptions about their spending and and look at their look at their stated returns. And it gives you an idea of magnitude and direction. It's not going to be precise and and you know, Amazon would be the better example, I think, because AWS, you know, very early on was unprofitable for a number of years until it hit its stride about 10 years ago. And so so the bet might be a correct one. We'll just have to see and and so what we do, what we're looking at right now in in the analytics I just discussed is what is the increase in adjusted operating income year-over-year and then annualize that and then what is the increase in the capital base year-over-year and and as the denominator, the increase in operating income annualized would be your numerator and and how is that trending? Is it going up or is it going down? And so, for example, last night Oracle reported and their incremental increase in operating income divided by their incremental increase in capital year-over-year was about 8.5%. That's below their weighted average cost of capital. It's positive, but basically it's it's telling you so far they're destroying value. Now, for Microsoft, that same number is almost 40%. So, it gives you a pretty good idea of of of so far who's kind of winning in this and and who's, you know, spending a lot of money for not a lot of return. But again, like Amazon, that can change. So, we'll just have to see and keep monitoring it. As I like to say, predicting the future is hard. So, um, people are doing it. The difference now is is they're committing a lot of dollars to these projects about predicting the future and that's a new risk.
Right. And 8.5% it is positive, as you say, but doesn't sound terribly high. I think the, you know, long-term Oracle bonds are have a 6.5%. And so there is a spread to harvest there, but I'm sure with cost of equity and stuff it's...
With cost of equity, it's above that. Yeah.
Yeah. Um, and that is what the return is for Oracle right now where things are white hot and all of this money is being raised in venture capital to back companies like OpenAI and Anthropic that are the customers of of these things, which is why the the money is so good right now.
So so it could it could look worse when things get a little when things cool down a little bit.
Yeah. So that's the that's the that's the sort of left side of the tail, right? Is that that these returns are all based on on in many cases selling to unprofitable companies and and as long as as those unprofitable companies can keep raising money to keep paying those bills, that's fine. But if we get a credit crunch or we get a 2001, 2002 pullback in sentiment, um, you're going to see a lot of that spending drop and drop pretty quickly. So it's just again, it's just added risk. The risk levels for all these companies are much higher than they were now five or 10 years ago.
So of the giant companies in the S&P 500, it sounds like the ones you have the most concerns about are Oracle, Amazon, and Meta. If you had to rank those three from least concerned to most concerned, which would it be?
No, we would we would we would be we're most concerned right now about Oracle in that whole group and and um that that's pretty much our only negative view on the whole group is is Oracle because it just their their spend plan is just so enormous um relative to returns. Amazon has a lot of other businesses and we think that our analysis of Amazon is is skewed a little bit and and the rest in effect, as I say, we're long through our hedges. So the only the one of the only one of the hyperscalers of the mega caps that we would be concerned about right now is Oracle. But but we're short lots of others, I say, secondary and tertiary plays.
And yeah, Oracle stock over the past few months has been suffering and its credit default swap for five year has has widened out significantly. How much of this do you think is a potential, you you earlier said bet the company, you know, a potential existential question because, okay, you know, if Google or Microsoft is going to waste hundreds of billions of dollars, like ultimately they will be okay, but Oracle is borrowing all this money and it's, you know, it's investment grade, but it's, you know, for a long time hovered on that investment grade kind of line, you know, I mean, do you think that there could be some substantial credit issues for Oracle?
I mean, if if AI if AI monetization gets pushed out and if it's not a 2027, 2028 um, you know, occurrence but 2030 or or never, then Oracle will have will have, you know, fundamental financial problems. It's not at that level yet, but certainly certainly they will they will be under a lot of stress if they continue to borrow money to do their buildout as opposed to use equity. I have to think at some point they might they might have to reconsider and and start issuing equity to keep the balance sheet in check. So far, it's been mostly done with that.
And what fueled the the surge in Oracle shares during the second quarter in the early summer before the, you know, quite drastic correction we've had? What fueled that surge was I think it reported earnings and its remaining performance obligations surged from 137 billion to 455 billion.
Yes. Which some people think of as, oh, this is just, you know, future demand and this is what's going to be revenue and earnings in the future, so it indicates incredible demand. What do you think about that number? How concentrated is it among one customer, mainly OpenAI?
Yeah. Yeah. It's not a gap metric. It's not a real backlog. It's, you know, it's if everything goes right and they they continue to to deliver, you know, our revenue should be this over the life of the the contracts. I mean, again, I I the stock reacted positively to it, but again, this data center and and hyperscaler and AI stocks were all reacting in the second quarter to those kinds of announcements. It wasn't just Oracle. I should make that clear. I mean, any company that that came out and said, oh, we're going to build data centers, went up 50 to 100% during that period. So I I I look a sconce at at the RPO number and and I think investors should give it a wide berth.
All right. Now let's talk about the, you know, lesser known smaller companies, the so-called Neoclouds. These are CoreWeave, Nebius, former Bitcoin miners like Iron. Tell us what role do these sorts of companies play in the ecosystem and why is your concern and skepticism concentrated in this area?
Because again, these are these are the guys that are hosting the GPUs, not not benefiting from the GPU output. So it gets back to my my first comment to you that that we think the money and the magic is going to be made from what the chips produce, not where they sit. And and I just think that that hosting hosting GPUs or owning them and renting them out to to third parties is just not a particularly good business. It's a commodity business and it it these are landlords. They're not tech companies and I think that that's that's the basic difference here. And with everybody rushing to build out data centers, you know, it's a supply and demand story, right? You're going to have more and more people who will be having facilities. There seems to be this canard out there that that well, if you have land and power, you have a monopoly or you have an oligopolistic situation where you can charge, you know, ridiculous rents and and nobody else can can get them and all you need to do is look at Iron's Microsoft deal, which was trumpeted at the time as as this great deal and if you sort of drill down and analyze the returns, it was a mid-single-digit return on capital deal. And and so, you know, if that's what a monopoly type type provider gets because they have access to power and land right now from a really good credit, I mean, you know, yawn, who cares? And and so I think that that's that's the core of our our negative view on these Neoclouds and Bitcoin miners is that they're all jumping into a business that's basically commodity, low return, ultimately business.
Tell us. So you're very familiar with the old school data center stocks. Um, I don't think you mentioned them, but you know, they they might be stocks such as Equinix, such as Digital Realty Trust. You said earlier that they they capitalize expenses that they should go through the operating line and they overall earn low returns. So how would you estimate overall, like the returns of that business? you know, not the stock price, but just the overall, like economic returns of those two companies, publicly traded companies, as well as the, you know, the privately traded, you know, private equity landlords that, you know, are quite big in this business too.
Yeah. So, so again, the the Digital Realities and the Equinix's are just colocation businesses, basically. Um, Equinix has an interconnected business, but but for the most part, these are these are just simply companies that provide the real estate, provide the backup power, provide the the power itself, the cooling, you know, what have you. And you could just look at at the stated returns and that's based on sort of 15 to 20 year depreciable life for the assets. And using that figure, their returns are really, really low. They're they're low single digits for Digital Realty, mid-single digits pre-tax for Equinix. But what what really gets interesting and what the reason why we're still short them is if you actually look at the cash flow statement, you'll see massive amounts of capex that continue and it's not new builds. Most of it is is for the existing data centers, but they have this little accounting game they play because they're REITs where they claim that that 90% of that spend is growth capex and not recurring capex and and everybody sort of says, see, oh, there, you know, don't count that capex because it's for growth. Well, the accountants let them call it growth capex. If it's in an existing system and you can either acquire a new client or raise prices on existing clients, then you can capitalize, you can categorize it as growth capex, not recurring capex. But if the HVAC system goes down at a data center, it goes down. You have to replace it. How you categorize it, you know, is sort of semantics. And if you can persuade, you know, your auditors that, well, it's we'll be able to raise prices and maybe get one new customer by putting in a new HVAC system, you know, it's growth capex. Meanwhile, these companies aren't really growing. I mean, they're growing revenues mid-single digits. They have 20% vacancy rates. It it's ridiculous. And and they they both Digital Realty and Equinix have capex greater than EBITDA. So they're financially stretched as well. So they're in effect borrowing money and issuing stock to pay interest and dividends. And that's just never a good position to be in.
And is the reason that these companies kind of exist because the large companies, you know, Microsoft and the Microsofts of the world basically don't want to have this on their balance sheet?
I think that's a really good point. Yes. The answer is is that it's excess capacity and in the in the cloud world, you might have the right geography because of the need for clients to access data quickly, latency. So so that does matter more so than in in the AI data center business, for example. However, you raise a really good point and that is is that increasingly we see with Microsoft's announcements, with the Meta announcement and Blue Owl and their financing, more and more of the hyperscalers want to get these assets off their balance sheet and lease the capacity from somebody else and let somebody else take the actual, you know, capital intensity risk. Um, and I think that's a really interesting interesting point that we've made to clients is note now how many of these companies are trying to not have this stuff on their balance sheet.
So let let's now talk about CoreWeave, which I think is the biggest and the most indebted of these so-called Neoclouds. At at CoreWeave, is their basis their business is owning the data center or what what else is is their business other than that?
Their their their business is not owning the data center. They're like a virtual. They're they're leasing the data centers from others and owning the GPUs.
And then selling that compute to the clouds?
Yeah. Yeah.
To the hyperscalers. Yeah.
All right. And what do you think about the returns on those businesses?
Yeah. So that that's a bet strictly if you're just basically in the in the GPU middleman, if you will, you know, that's just a bet on on depreciable lives. And and I think that that's that's what you need to, you know, that's the that's the bet that that you're buying these things and leasing them out and that five or six years is is is actually too conservative, that you'll be able to earn money on them for 10 years or 12 years. And and that's the bet you have to make if you're if you're a CoreWeave investor. The bet there is that those chips last for a long time so that every year you're not depreciating that much. If it's if the chips last for 10 years, you only depreciate 10% as a cost every year. But if they last for three years, you you have to depreciate a third of what you spend, which is a lot.
Right. Right. And and and again, the question gets back to your comment, and that is if these chips last so long and Microsoft is is depreciating its chips over six years. If these things last longer than six years, why is Microsoft leasing the capacity from you?
That's a good question. So explain what is the depreciable lives that the CoreWeave, the Microsoft of the world are putting on the chips that they're buying and then expensing, and what do you think is a more appropriate depreciable life?
We're using somewhat in our modeling, we're using something similar to them. So they're using six years, as I recall, and a number of the hyperscalers really increased that lately from five years to six years. We're using five years with a 20% residual value. So the numbers, you know, are are somewhat similar ballpark and and and again, I I'm just looking at rental rates for GPUs and other things to get some third-party verification on this. You know, there's an there's an index on Bloomberg quoting spot rental prices. I think it's down 28% year-over-year currently.
That's the Hopper GPU rental index.
Yeah. Yeah. And and and you know, that's one one check. You have some of the chip companies themselves, CEOs have said five years, six years seems right. People believe that if you use them 365 days, 24 hours a day, that the physical life is not much more than six or seven years. They only last 10 or 12 if you're, you know, not using them um at 80 to 100% capacity. So, we'll see. We're going to again, it's one of these things where where we're going to see. I would not want to be betting my company on, you know, 8, 10, 12 year life on these things. I think that that's that's an asymmetric bet that that you're going to lose on. And conversely, if it's two or three years, then all these companies are in trouble. So, we'll have to see. I you know, it's it's something obviously we and other investors are monitoring pretty closely. But again, that that's that's the bet there that that simply, you know, they're smarter than everyone else and have figured out that these GPUs have incredibly long lives. And they said something on their last call about, well, we're still, you know, renting some of these chips out after the term is over and at at basically 90, 95% of the last deal. And I I'm I'm highly skeptical of some of those comments, whether they're cherrypicked or whatever, but there'll be enough GPUs out in the marketplace, you know, within a year or two that we'll have a pretty good idea of what rental prices are.
Right. I don't really understand the technology at all, but Nvidia is just such a a a dynamo in terms of improving their architecture basically every year. So there was Hopper and then this year we have Blackwell from Nvidia. Next year we're going to have Rubin, maybe maybe the next year if that, we're going to have Fineman. What you were talking about on on Bloomberg is the Bloomberg Hopper GPU rental index. So this is the old Nvidia GPUs and it shows that that the cost to rent these has gone down by so much. So what what you're saying is does not in any way contradict the very rosy outlook of Nvidia is transforming the world with AI. You I mean, it sounds like you kind of agree with that, but because they are it's so transformative and year after year, just the old GPUs are not that exciting and valuable anymore and therefore they the people who buy those should depreciate it at a faster rate and a higher depreciation cost. That's what you're saying.
Right. And and and again, you know, when when CoreWeave or others say, well, we'll we'll simply use the old chips for other tasks like inference and not training. That's great, but that should reflect itself in rental rates and and so we will have a check on this in by the free market as to to how valuable these chips really are, you know, three years out, five years out. So far, we're seeing we're seeing depreciation, you know, reduction in rents consistent with with sort of four to six year life.
Talk about how indebted CoreWeave is. How is it financing this incredibly capital intensive growth?
So they they they're doing what the Bitcoin miners are. They're using lots of converts. And I I kind of laugh because when I see people throwing throwing bricks at me on on X for my comments, I often see the same people saying, "Oh my God, they just did a 1% convert deal. It's like free capital." And and not not understanding that converts for for some companies can be really, really expensive when you factor in the option value that they're giving up. A convert for your listeners is is composed of really two components, right? A a call option based at some premium price, the conversion price for typically five years, seven years, whatever the term of the the bond is. And then the remaining part is in effect a very low coupon or zero coupon bond that will mature at par at at maturity date. So you can you can kind of once you get the option value of what you've given away and the convert, you can figure out the the yield on these things and you know, they're they're they're quite high. The cost of capital is quite high when you factor that in. But retail investors seem to think that the opposite, that this is just 1% cost of capital because that's the coupon on the bond.
And if you knew for a fact that the stock would never hit the strike price, maybe it would be a 1% cost of capital. But if the strike, if it hits that strike price, there's going to be a very large dilution, which is basically just issuing shares.
Yeah. Yeah. And we just saw that with Iron. They just did a convert at current prices to buy out the converts that it had issued like a year, 18 months ago. And and they paid I think 3x from from to buy out those bonds that they issued, you know, less than two years ago. And so in effect, those bonds which would have been seen by the same people as cheap capital, you know, ended them ended up costing them quite a bit in terms of new share dilution and and so there's no free lunch on this and and the fact that so many of these companies are trying to issue converts is in effect it's a backdoor way of issuing equity.
And what do you think stops kind of this this train, this circular reflexive loop where the more capital is raised, the higher the revenues go, and therefore the higher the revenues go, the more capital can be raised? What, you know, you've been on Wall Street a long time, you study financial history. What stops this circle?
You you never know. I mean, you know, in in 2000, we we can look at a couple of things that that occurred that that changed perceptions. But but the negative fundamentals in in the dot and the telecom boom didn't start showing up till '01, even though the stocks began going down in March of 2000. So the stock market correctly anticipated the the decline in orders and earnings that occurred in '01 and '02. And and if I had to look back and and one of the things on the my reading list for my course is the wonderful work that Anthony Adzilo at University of Minnesota did in the late '90s and early 2000s when he basically punctured the prevailing wisdom at the time that internet traffic was doubling every quarter. And that was something that MCI and others had, WorldCom had had told people and it it kind of got out there as is the accepted wisdom that internet traffic was doubling every three months. So people were ordering like crazy and it wasn't just the fiber optic companies. I want to dispel a myth. Most of the ordering during the telecom bubble was by regular corporations building out their networks or very profitable telecom companies like local local bell companies, the other long-distance companies. Fiber optic companies were a very small part of that. And so everybody basically cut back their orders when they realized that well, maybe, you know, we don't need all this equipment in 2002 and 2003 if the growth isn't going to be 16x a year, because that's what it is if the internet traffic doubles every three months. It's up 16 fold a year. What Zilko's work showed was that it was doubling every year. So so up 2x, not up 16x. And and that makes a world of difference for your ordering if you suddenly realize that you only need 1/8 of what you thought you were going to need. And and so order books collapsed in late 2000, early 2001. And you know, with a very mild recession, very mild in '01, '02, we saw corporate profits drop 40%. And and and so it's and now, if you think about how much of economic growth and corporate profitability in the tech sector is based on data center buildout and AI, you easily could see people, you know, pausing their spending. And in any kind of capex boom that translates immediately to to earnings. So if demand is only growing at two times a year, it's doubling every year. That sounds like a lot. But if if the capex is 4x a year, that's too much. And so ultimately it it slows down.
Is that how you see this playing out? Is basically dot 2.0?
I I I don't know. I mean, again, that's what happened the last time and it just and it blindsided everybody and and so, you know, nobody at the time was saying, yeah, we got a big order book and our revenues are growing dramatically, but it's based on, you know, it's based on a really optimistic view on current demand and if that drops, you know, we could be in trouble. No one was saying that and nobody's saying it now. I'm just pointing out that that when anytime you get a capital spending boom in a certain area driving your whole economy or driving your corporate sector profitability, and increasingly we have that now, then you run the risk of people overestimating their need for all that capital equipment and capital equipment, unlike, you know, regular consumer consumption items, can be switched on and off, you know, abruptly. We don't need eight data centers, we just need three. And and so as opposed to, I'm going to buy less hamburgers this month, you know, and so I think that that's the amount of earnings risk that is coming from this is increasing rapidly and does look a lot like '99 in 2000.
And I want to ask you about that cyclical nature in just a bit, but you because you mentioned that that paper that I'll have to check out showing, oh, it's only doubling, uh, uh, twice a year, not 16x. What do you think about the claims right now about how, you know, ChatGPT and the large language models are growing faster than the internet? They're going faster than everything else. What do you make of those claims and and that data?
Yeah. Well, I mean, again, we'll have to see. I mean, some of them are coming off a pretty low base, you know. So, so it's not surprising that maybe from 2022 or 2023, it it gets a lot harder to keep doing that when, you know, when your revenues are 50 or 100 billion. And so what well, that's why I keep pointing out to, you know, this 2027, 2028 period is going to be so crucial because all of these companies are going to be monstrously big in terms of their balance sheets and at that point, the growth has to continue because if it doesn't, they're not going to be able to to uh either keep ordering or service their debt.
Are you a sophisticated investor concerned about today's markets? With the ever-present potential for stock market corrections and the risk that bonds won't provide the hedge that they used to, you may want to consider adding an asset class that has historically offered stability when you need it most. The prices of agricultural commodities such as corn, wheat, soybeans, and sugar typically move to their own rhythm driven by factors like weather and geopolitics, not the factors that move more traditional assets like stocks and bonds. Since inception in 2012, every time the S&P 500 has had a correction of 10% or more, the Tukrium Agricultural Fund benchmark index has outperformed. Getting exposure to agricultural commodities has always been tough for individual investors. And while it's not possible to invest directly in this index, Tukim offers ETFs designed to give investors price exposure to these markets. If you're a professional investment adviser or a self-directed investor managing your family's wealth, click on the link below to learn more about how incorporating agriculture into your portfolio can help diversification and protection. Thanks for listening. Let's get back to today's interview.
That nature of double ordering and just, you know, pounding the table to buy as much as you can in a boom and then totally pausing and canceling orders during the bust. You know, we could criticize that, but in some sense, it could be rational because if you're an order manager, like you you need it and to or to make sure that you just need that you just get what you need, maybe you have to order three times what you actually need. It can be rational. What what do you make of the argument, Jim, that that you know, particularly the semiconductor supply chain that was characterized by this boom and bust dynamics, that it's right now in AI, it is different because that the growth is secular and, you know, you're not seeing the the iPhone cycle or the cycle that, you know, that is kind of like this. It has been a super cycle, maybe, but may maybe doesn't continue. What do you think?
The real the real boom bust in dot wasn't semiconductor, it was telecom equipment and that's where the bust occurred. And so I again, I I think that where we're seeing it now is in the massive massive capex going into AI and data centers and and um that's clearly where it is. I mean, you can't can't miss the numbers. So we're just going to have to see. And by the way, as I would say, most of the most of the companies ordering that telecom equipment were corporate America, right? Building out their networks and for the internet and and and they were profitable. Customers were profitable. They just cut back on their their their spend. What's scary about this one is that the underlying customers that are spending a lot of the money are unprofitable. And and and so it raises it raises the risk level, I think, in a way that's worse than 1999, 2000. Again, the the only core unprofitable entities back then were the fiber optic companies and the CLECs, the so-called competitive local exchanges, which were local local phone, you know, virtual phone companies that were popping up in every city. But the combined capex of fiber optic companies and CLECs, the two unprofitable business models in the telecom bubble, was a hundred billion dollars from '97 to to '01. So it was it was about whatever $20 billion a year in capex from those two unprofitable business models. And think about think about the unprofitable AI companies and what they're spending today. It's a lot more than $20 billion. And so it's a you have you have a much riskier construction of the demand than you did in the telecom and .com era.
Okay. I'm glad you brought this up, Jim, because I have said and it has been said many times on other financial media, the argument that, oh, right now the people who are buying the Nvidia chips are profitable companies, Microsoft, Amazon, Meta. And I stand by that statement in what it is actually who actually is buying the chips. You're saying the underlying customer, the people who are buying the compute from Amazon and Microsoft, and the reason why Amazon and Microsoft are buying the chips in the first place are OpenAI, Anthropic, and other unprofitable...
A big a big chunk of them are and not all of them, but but yeah, but a very a very large amount of that that underlying demand and that's why I keep getting back to you on this '27, '28 you know, timeframe, are are inherently quite unprofitable. And so we'll have to see. The the unprofitable companies as a percent of the spend is higher in this cycle than it was in the telecom cycle. I guess that's my message.
That is really important, Jim, because I think a lot of the market believes that now, oh, it's what I said, it's Microsoft, it's Amazon, it's the most profitable companies that have ever existed to to to spend. And they view view the .com bust, you know, like I did. Maybe they were maybe they weren't around then, but as, oh, it was the pets.com that are spending the money. You're saying you're saying it actually was those companies were tiny. I mean, the the companies everybody sort of laughs about in in the IPO boom of the dotcom era were tiny. The real spend was coming from the General Electrics, the Coca-Colas, the AT&Ts, the the local, you know, the telecom US West and PacBell and and Verizon. Those companies were immensely profitable and and they were the ones buying from Cisco, Lucent, Nortel, and and then cut their orders. Again, the the inherently unprofitable companies that everyone points to, the fiber optic companies and in our experience, the CLECs, which all went bankrupt, they were they were relatively small part of that spend. And and you know, we lived it, we we were short these companies, so I remember it vividly. But most of the telecom spend that got cancelled was and and network spend was from very profitable Fortune 500 companies.
Wow, that's uh that is very, very interesting. Um, I'm glad you tell me that. Jim, uh, you said you spoke somewhat favorably of OpenAI. Because in the the com the companies that benefited from all the the spend was the the Amazons and the Microsoft that actually did reap the the benefits of all this technology. I wonder, Jim, if OpenAI was publicly traded, you're telling me you wouldn't take a look at it on the short side, given that HSBC forecasts half a trillion dollars in operating losses?
If they go if they go public, we'll certainly take a look at them. Right now, right now, it's all forecasts and, you know, and whatever Sam Altman says it is, but we'll have to see. Yeah, I mean, everything you you said about Iron, CoreWeave, I I think in terms of scale, it is, you know, OpenAI is the king right now of losing money and of of uh spending a lot of operating uh of uh uh capital expenditures.
But if if AI is truly truly uh the the technology that its proponents believe it is, then it's most likely that that those companies will become inherently profitable or monopolistic, much like Microsoft was during, you know, the the CPU era. And and so, you know, they'll they'll turn out to be good investments. We'll we'll have to see. We don't know. Um, they're claiming they will be able to monetize all this stuff. There are skeptics about that. There are believers about that. What I can point to now is that the the guys building the infrastructure around that are the ones building bad businesses. But OpenAI is kind of at the at the heart of it of of their spending the money. I mean, how over the next five years, you know, I mean, HSBC forecasts that their revenue could get to over 200 billion, which sounds great to me. Um, but again, they have 500 billion in operating losses, which includes uh depreciation on the what what the most rosy OpenAI bulls and Sam Altman say to the HSBC forecast, as well as even more bearish forecasts of they don't even get that revenue that high. You know, the total AI bears, where are you on that spectrum? Would you say or how are you thinking about it?
Again, I don't have to be anywhere on that spectrum. I I can look at it and and and look at the hyperscalers on one side and the and the the data center companies, Neoclouds on the other. I don't have to have an opinion yet on on that. We'll have to see. I you know, there there needs to be concrete steps to monetization for these companies sooner rather than later, I think. But, you know, I I my prediction about what happens in 2028 is no better than anybody else's.
Okay. But something maybe you do have a view on is how is this being funded? So in the publicly traded, it's funding in the debt markets as well as convertible bonds. OpenAI is raising a lot of venture capital. Do you have a sense on, you know, just how capital intensive it is? Like how much money it needs to raise next year? I think literally this morning we got the headline that Disney is going to invest $1 billion in OpenAI. You know, uh, I remember back in the day when a billion dollars was a lot of money, but that is not going to fill fill this bucket. And, you know, I mean, what are they going to every week they fly over to Saudi Arabia to get another billion dollars? Like, how how is this going to play out?
Well, I I I, you know, you'll have to talk to the venture capitalists. I mean, these are enormous numbers and everybody says the dry powder is there and eager to invest, and it is until it isn't. And I I just, you know, again, it was there to build out the telecom stuff until it wasn't. And venture capital dried up almost instantaneously, you know, in the third quarter of 2000. It's it's it's very it's very interesting and and to use George Soros's term, it's maybe reflexive. If the stock prices of these entities start going down, you know, you could also see the
The capital drying up as opposed to people wanting to invest more at lower prices. This is the kind of situation where they were investing more at higher prices and the purse strings could literally close up if stock prices in this area begin to decline. So there's a lot of it is a confidence game.
Jim, it's so important. Earlier you said that the spending in the dot telecom boom ended at the end of 2000. You know we know that the peak in stock prices was the beginning in 2000. So a lot of people, myself included, have this mental model of the price follows the fundamentals and the fundamentals change stops first and then the price goes down. But actually you're saying the price the market decides first.
The mar NASDAQ I believe was down 30% from early March to mid-April in 2000 on no in effect no news. It was before even March quarter earnings came out. The market just started going down and what the market correctly at least in that selloff was predicting was that order books were about to you know collapse. Now that didn't happen for another six months but the market kind of figured it out before even the purchasing managers did. And so if you're waiting for some of that that kind of evidence to show up, I'm just saying be mindful that there could be a big impairment of your capital happening before that even happens. And then you'll see the verification of why it dropped later. And that's something I think novice investors don't understand. And there are plenty new novice investors in the market that weren't around in 2000 that I think, you know, might be surprised if that happens. That's another really important point.
What do you think about Nvidia, Jim?
They have a good business. I, you know, I the amount of vendor financing that they're doing is so far not material. I don't know why they're doing it, i.e. investing in their customers and doing some of the same stuff that Lucent and Nortell were doing back in 99 and 2000. Um, they don't need to do it if demand is as good as they say it is. You know, they could sell all these chips twice over. So, I'm a little perplexed at some of the things that they're doing, but it's a cash machine currently. They're the biggest player in town until Google, you know, and Amazon maybe and AMD begin to put out their chips to compete. But, you know, right now it's there are much better places for skeptics and short sellers like me to be looking than Nvidia.
Mhm. What about Palunteer?
No opinion.
Jim, tell us about the alternative investment industry, private equity, real estate, and increasingly private debt that is now getting involved in this game. You know blue alowl as you mentioned earlier did an off-balance sheet special purpose vehicle to finance Meta's data centers. What do you think of that business and as well as private credit in general which has just exploded in growth?
So I've been kind of skeptical about private equity for a number of years now. You know, I pointed out in the investment committees I sit on and to clients that private equity was basically leveraged leveraged equity and that returns to private equity, while good, were not commensurate with the risks you were taking. And that the whole idea of marking your portfolio sort of arbitrarily was hiding the risks that in effect you were buying leverage, small and midcaps and it levered 2:1 or 3:1. Well, then your return should be a hell of a lot better than 10 to 15%. And that's what private equity was returning when the public markets were returning 10%. And you know, I said sooner or later the public markets may end up doing better than private equity. And I think now that's the case for the last five years because the fees structure is so huge and in effect a lot of the kinds of companies they were buying at the time have not done as well as areas like technology.
But today my concern would be the sales job being done on private credit because again, being a bit of a gray beard dinosaur, I've seen this before. The concept of private credit is that you are going to earn equity rates of return for basically incurring senior level debt. That is amazing, right? To be a debt holder and earn equity rates of return should not exist in a perfectly efficient market. And through great underwriting and good analysis, all these private credit companies are telling you that they can give you 10 to 15% returns for owning in effect senior or senior secured paper. Now, everyone kind of forgets because a lot of people aren't as old as me, is we heard this before. And when did we hear it? We heard it from a guy named Mike Milken in the late 80s who told us that junk bonds had an excess rate of return, an equity rate of return because they were lower rated and people shunned them. And if you took the risk to buy these bonds, you could earn equity rates of return. And even if there were defaults, the recoveries would be so great that they would offset any default risk. And so therefore, you were protected.
And it rested on two false assumptions that people didn't realize until Ed Altman and others in 1989 pointed it out. Number one, that most of the excess returns that Milken pointed to from the 50s and 60s and 70s in lower rated debt were from so-called fallen angels, companies that had been investment grade but cyclically became junk and then the cycle turned, they became investment grade and thus bonds went up and that was your return. Number one, they were not from new issuance of junk. And number two, that as time went on in the 70s and 80s, Milken and Drexel was pointing out the default in any given year based on the total amount of bonds outstanding at during that year. Which of course is not the analysis you should do because if the market is growing rapidly, lots of new companies are issuing debt. The denominator of course is going to grow faster because those bonds haven't seasoned. And then what you really needed to do was look at the bonds that you issued in 1979 relative to all the bonds that were there in 1979 and follow it on a cohort basis. And when you did that, you saw that the default rates were not lower and the recovery rates were not higher than people expected. And that adjusted for that, you were getting a corporate debt rate of return, not an equity rate of return. And that whole thing kind of came undone in the real estate and S&L bust in the late 80s and early 90s.
There's one other thing though that Milken did that was genius at the time and we noticed it and now we're seeing it again in private credit. Increasingly, he had his network of companies buying high yield bonds, own regulated entities like S&Ls, insurance companies or trust companies. And because regulators would prevent the owners of those companies dividending up too much money to keep the regulated entities safe and sound, what he did was he had those entities buy the junk bonds of other companies in the network. And then when your parent company needed to raise money, you'd issue junk debt and everybody else's regulated S&Ls and insurance companies would buy the debt. Well, we're seeing an echo of that with more and more of the big giant private equity credit companies increasingly own regulated entities like insurance companies to buy this debt. Apollo has a theme for example. And that's something that I think is worth keeping an eye on because you have increasingly retirees who are the beneficiaries of annuities from these companies increasingly being buying a lot of this debt and taking maybe risks that they don't realize they're taking.
And Jim, what Apollo might say and has said is, okay Jim, sure, but over 90% of the assets that we own through Athen are investment grade.
Well, that's certainly fine, but that was also the case of a lot of the regulated entities back in 1989 that went bust doing this because they were leveraged 10 to 1 or 20 to 1 like Executive First Executive, Columbia Savings alone and a number of companies that were stuffed with Drexel junk. It just depends how levered your equity is, your statutory capital to your assets. You could have 90% investment grade, but it might still be 100% of your equity is in that stuff.
Yeah. And particularly the thought of a private credit loan being investment grade, does that give you pause?
Well, it depends. I mean, again, there's I'm of the view that capital markets are pretty efficient and to say that this giant amount of credit out there is just inefficiently priced so that the investors in it can earn 10 to 15% when often the coupons aren't even 10 to 15%, right? They're seven to eight to nine to 10. And so there's some inherent leverage in the funds going on to get you to that higher rate of return. And then of course, we get into the whole thing about shopping for ratings and how valid are some of these third-party work that's being done on the credit, but that's a separate issue. And they make the 15% by buying a 9% yielding bond, levering it up slightly. They may also generate that 15% historically by the 9% loan narrows because there's this huge credit boom at 7% so there's a mark-to-market gain on that.
Yeah, it could be. I mean, there's always nice innovative ways for people to goose returns, but at the end of the day, it just usually entails taking more risk.
Yes. And Jim, I want to ask you about so historical LP limited partner returns for private equity and private credit have been quite good on paper. It's also true that investing in the publicly traded GP general partner like Apollo or Aries, you have done very well in that business. I've tried to look through their accounting. What do you make of the Aries, you know, and Apollo Blackstone just their accounting? I'm not saying that there's anything wrong with it at all, but it's extremely complicated. You know, even to someone like you, it must, you must be scratching your head at least once, right?
Yeah. We haven't looked at them recently, but the answer is yes. They're complicated, but typically they turned out to be much better investments than the LPs gotten their deals. And I think that's an interesting observation that you should have invested in the GP, not the LP. But no, you're right. They're incredibly complicated. Take a look at something like Brookfield, which is just a maze of interconnectedness. But we haven't looked at these companies recently.
Jim, how would you say this bull market right now compares to the bull market of maybe late 2021? I mean, we said after 2021 that was the most speculative market I've seen in my career, just that sort of six-month period post GameStop because you had the meme stocks, you had NFTs, you had a big rally in crypto. Um, you know, lots of things going on there. Um, you know, SPACs at one point SPACs were raising two to three billion a night in February of 01, which was equal to the entire US savings rate, so you knew that wasn't going to last. Um, I don't there were periods in 2025 it gave 2021 a run for its money. Let's just put it that way. Um, you know, I certainly this late spring and summer with the run of AI and nuclear stocks and quantum computing stocks, you saw a lot of the same kind of behavior you saw in the first six months of 2021. You didn't get the issuance that we saw in 2021. So that's a difference. But things are starting from a higher level. So you know that's the offset to that. And then you had a big move in crypto and you had the crypto treasury companies which was absurd. So there were some parallels.
Yeah. A very good short by you shorting MicroStrategy and going long Bitcoin against it to harvest that premium. Jim, why do you think we haven't had the issuance this time around unlike 2021?
I think because a lot of it's happening in the private markets. So the private markets are so much bigger than they were in 99 and 2000 that they're willing to finance lots of things and keep them on their books longer than go public. In fact, ironically, if you go public, you may find a valuation lower than what you've been marking at in your private fund. So there's a little bit of a catch-22 at work, I think too. But I think it's just because the advent of the private markets is so much greater than it was 25 years ago. The private markets can handle a lot of the capital needs of these companies.
You know, Jim, I've been following you for a long time like a lot of people. And I really liked your explanation of why for a time, I probably not true, I don't know anymore, but you were short Uber expressing very skepticism about Uber. It's just crazy to me like for how long your thesis was right of the fundamental economics of the business. But it just does appear that maybe Uber has kind of reached escape velocity and now is somewhat profitable. But
Yeah, you were right for a really long time. We are no longer haven't been negative on Uber for a few years now because they did reach escape velocity, but of course the valuations in the early years was just way too high and people paid too much in the IPO and post-IPO.
What else is drawing your skepticism in the public markets right now?
There's a lot of things that are basically, you know, we're looking at in terms of areas. We've covered a lot of them today. Um, but we have a lot of sort of one-offs. You mentioned Live Nation that we like that we don't talk a lot about publicly that are just really interesting. I mean, often accounting stories. Um, we published something on and I went online to talk a little bit about one of them earlier in the year, Erie Insurance, which was playing big accounting games with its insurance subsidiaries where it basically fobbed off its insurance company to the policyholders so it could take that off the books and then just charged that in those insurance companies a flat fee of 25% of their premiums and held themselves out to be a service company. Well, the underlying insurance companies had no employees, no directors. Erie, the publicly traded company was doing all that for them. So really, in terms of our view, it was not arms-length. And Erie's only customer was the Erie insurance companies and vice versa. The Erie insurance companies relied totally on the parent's employees to do their business. So this was simply an accounting dodge to get the relatively mundane and recently unprofitable property casualty business off their books and just show a service income off the top line. Um, and that stock's come down now quite a bit this year as people have come around to our point of view on that. So those are the kinds of things that's kind of our bread and butter that we still look at for our clients and, you know, the data center stuff is fun and interesting and timely, but you know, we sort of thrive on the ideas that are more idiosyncratic and one-off like Live Nation or Erie or what have you.
We'll leave it there. Jim, thanks so much for coming on. People can find you on Twitter, realjimchainos. Thank you everyone for watching. Please leave a rating and review for Monetary Matters on Apple Podcast and Spotify and subscribe to the Monetary Matters YouTube channel. Thanks for watching. Remember to click the link in the description to learn more about the Tukrium Agricultural Fund benchmark index. Until next time. Thank you. Just close the door.