Full Transcript: Andy Constan on Yield Suppression and AI Funding

Justin: Andy, it’s good to see you again.

Andy: It’s great to see you guys. Hey, Justin.

Justin: These episodes with you that we do on a monthly basis, I think have been resonating a lot with our listeners and our subscribers on YouTube. And I think what investors sort of appreciate when we’re able to sit down and talk with you is you can kind of take these complexities that you’re seeing in the market, that we’re seeing in the economy and all over the world, and a lot of different things, and really try to think about those from a first principles perspective. And you have a good way of kind of getting underneath the surface at the key things that are driving markets and the economy and stocks and everything that you’re sort of always looking at for your subscribers on an ongoing daily basis.

And I think today we kind of want to focus the conversation on a few key ideas. The first is, and this is where we’ll start, where the market’s at today, given where stocks are, where interest rates are, and inflation, and kind of think through what investors should be thinking about as we’re sort of in this market regime and period of the market. And then we’ll kind of get into things like the risk premium suppression that you’ve been writing about, how Warsh and the Fed could change things up here. So a lot to talk about. Some of this comes from the thoughts you’re putting out there on Twitter. Some of it comes from the Damped Spring research you’re putting out there.

So always just appreciate you kind of sitting down with us and working through some of these things. So I guess, to start, the market, I think, has confused some investors in that here we are, stocks are at record highs. You have the 30-year near multi-decade highs. Inflation is still high. You obviously have the war ongoing in Iran. So how do you think investors should make sense of all of those things?

Andy: Right. I mean, I think markets are basically acting internally consistent at this stage. It’s not like there’s things that are saying, “Wow, that really makes no sense to me.”

Interest rates going up is often confused — I think even our president is confused about the idea that real growth, particularly productivity growth, also population growth, but that’s not a factor today. Real growth, by its very nature, pushes interest rates up because people prefer to spend their cash to build a data center that has great potential for earnings than buy bonds. And so the data centers are selling the equity and things like that. The investors are saying, “Well, I’d much rather own equity in that than own bonds.” And so that presses up real yields. And it’s consistent with an economy that’s running very hot. So I think that’s internally consistent in the rearview mirror. Higher interest rates, driven by higher growth expectations.

Down a level, interest rates are also a little higher because there’s a temporary period during which all this financing is being done, where there’s a new corporate bond issued every day, often many per day. The government continues to run a large deficit and continues to need to issue its own debt. And so there’s a little bit of what I would call risk premium expansion that’s going on that’s driving longer-term interest rates higher.

A lot of people, when thinking about the bond market, are making a lot of premature, in my view — but based on like five years of history, so sensible — assumptions that the Fed just doesn’t care about inflation. And the administration wants to run the economy hot. So there are people that are selling their bonds because credibility matters. I think that’s less of an issue than just very strong growth and quite a bit of bond supply.

And the equity market makes sense in the context of, interest rates are high, but they’re not restrictively high or not seriously high. The central bank isn’t really making... Even the most hawkish central banks are modestly hawkish. So there’s not a lot of restriction going on, and there’s strong growth, so equities do well in that environment.

Justin: Is there anything to be said for, obviously it costs us more to finance all of our debt now, and I think there’s some estimates that say, I don’t know what it is, but a couple years down the road, a lot of the money is gonna have to go towards higher rates, because we have so much debt. So is there a level — not in terms of the amount of debt, but a level of rates — that would kind of concern you more if we were to get there?

Andy: You know, for me, I don’t put much emphasis on that particular idea that we’re at a point where we are going to have a government-driven debt crisis because of lack of sustainability.

The way I think about that is, we have roughly $32 trillion of marketable debt, $8 trillion of Social Security debt, intra-government debt. You know, 32 trillion with bills at roughly 20% of that. The bills are already paying the interest that’s in the market. The rest of the capital structure is paying — the last I looked it was like 2.68%, but I think it’s slightly higher now. Maybe it’s close to 3% on average of what we pay. If that was 4% on $25 trillion of debt, it’s not nothing. $250 billion of higher interest. But it’s not gonna kill the economy to have that much additional interest paid by the government who can print the money.

So I’m not one that thinks that we’re on the verge of a financial crisis because of the rates the government’s gonna have to pay. I’m more focused on whether projects that are being financed by the private sector are going to outearn their cost of capital.

Justin: And I think that’s certainly something we’ll get to here. But I wanted to ask you — you mentioned inflation earlier. A few years ago, you wrote what you called the script to kill inflation. Can you explain what that script was?

Andy: Well, I think in most cycles, as the economy grows rapidly because of recovering from a recession, stimulus, lack of debt, all of those things, the economy cyclically starts an expansion. It ends with the private sector having a lot of debt. And when the private sector has a lot of debt, monetary policy can work very easily to cause a tightening, and that is by raising short-term interest rates.

This economy is very different from most economies, except possibly Japan ages ago, in that the private sector’s not gonna be particularly sensitive to the change in short-term interest rates. It’s not particularly levered, and in particular, it’s not levered to that rate. And so I think the short rate is not what is controlling the economy. I think the Fed, at least half the Fed, disagrees and thinks that’s the only tool that matters.

And so what I said is that what we have is not a traditional debt cycle. We started with an income cycle, that was driven by the government paying and financing with Treasury — financed by the Fed — a tremendous stimulus in 2021 and 2022 that flowed through to income, literally replaced people’s income while they were sitting out. And that was an entirely income-driven expansion.

Now we’re in a phase in which we have what I would call a wealth effect and a cost of capital for what is needed to be financed effect. Basically, anybody who wants to consume can dissave, because they have accumulated wealth due to the asset prices rally, or they could borrow — particularly if they’re a corporation, borrow in the capital markets at very low cost of equity capital and pretty low cost of debt capital. Credit spreads are pretty tight. And so saving and investment can be funded fairly cheaply for the last two to three years.

Now, of course, it hadn’t been necessary to fund because no one needed the money. There were no capital projects. People were staying in their homes. They were not moving. There was no relocation, so no new mortgages. And any private sector leverage had extremely low coupons because it was all done during very low interest rates. So the last couple of years have been driven by this dissaving and wealth effect.

And so to kill demand... I should have started with this. To get demand-driven monetary inflation back to target, you need to kill demand. That’s the only way to do it. I mean, I guess you could also deliver excess supply, and that’s part of the AI story that could possibly be disinflationary. But at the moment, there’s nothing like disinflation coming from AI, as we’ve seen by the price of our Apple phones that are gonna go through the roof because of the memory squeeze.

So for now, we don’t have a supply driver, so you have to cut demand. So how do you cut demand when short-term interest rates don’t work? The only way you do it is by hitting asset prices. And so that was what the script was about. It said to kill inflation permanently, you had to have stocks and bonds and gold and other assets go down, so the wealth effect would be eliminated.

We got there a couple of times. The fall of 2023, bonds and stocks were selling off pretty big until Halloween, when the policymakers decided that they had experienced too much pain and stopped the script. And so since then, we’ve had supply shocks, which are annoying and inflationary, and that has convinced some people — including the central bank — that inflation actually is coming back down because it’s just a supply shock. It’s tariffs, it’s oil, it’s Iran.

But I think many, and I believe this, believe that the supply shocks will affect inflation, of course, and once they’re resolved, that’ll be a downward pressure. But we’re not getting back to target with the cost of capital that is flowing through markets today, the wealth effect that’s flowing through markets today. So the script is you have to do that. And so I don’t know when that’ll ever happen or if the policymakers will allow it to happen, but because they won’t or because they might not, inflation is likely to stay above target for some time.

Jack: And I think this gets into the next question in your DSR, ‘cause the idea is, I think to do that, you want longer term interest rates up, right? And so right now you would argue the government’s sort of done the opposite of that. They’ve been suppressing long-term rates, both from the Fed side maybe and from the Treasury side. So can you talk about how they’ve been doing that and maybe why they’ve been doing that?

Andy: Sure. So the bond market’s a scary thing. Losing the long end of the bond market risks having inflation expectations unanchored, does have this modest financing issue you describe, but also flows through to the real economy and makes mortgages more expensive and so on.

So, well, gosh, it’s been decades now that the central banks have been unwilling to let the long end really trade to market. Risk premiums — that are a measure of how fair a deal a bond is, and for that matter any asset — have been compressed by policymakers at any sign of the bond market weakening.

The most notable one I guess would’ve been the SVB crisis. That was a handful of banks were gonna go out of business ‘cause they had unhedged extreme mark to market exposure on long-term bonds. The BTFP that was instituted to save those banks was orders of magnitude bigger than necessary. And that saved the long end. Nobody needed to sell any bonds. None of the banks that were in trouble needed to sell any bonds, and anybody who needed to sell bonds saw a fairly receptive market.

The Treasury has run high bills as their percentage of outstanding debt. That’s a protection of the bond market. Changes to the supplementary leverage ratio have supported bonds. Hey, the president even instructed the GSEs to buy $200 billion of mortgages just in the first quarter. That’s supportive of bonds because they financed it with short-term issuance.

And then any opportunity to cut the short-term rate by the Powell administration has been done. That was the fall of 2022. The stock market and the bond market bottomed because we got one soft inflation print and the committee said, “We’re done hiking. We’re done hiking. Please don’t worry anymore. We’re done hiking.”

So I have a number of those things. Those are some of them. And it’s a sensible idea. If you wanna keep financial stability as a real important goal, and avoid any sort of pain in the labor market is a goal, you wanna be easier, but the consequence is inflation lasts longer.

And so the Powell Fed was pretty willing to assume that, in fact, this wealth effect isn’t the driver for demand, that the only reason why we’ve had inflation is supply shocks, which would be transitory, and really expected inflation finally just to go to target without them having to inflict any pain on the economy. I don’t know that the Warsh Fed is any different than that, but that’s what we’ll have to see.

Jack: Well, they’re at least potentially talking like they might be a little bit different. And that’s one of the points you made in this piece — you’re starting to see maybe the drumbeat of this coming to an end. So what are you seeing that makes you think that?

Andy: Yeah. I mean, I guess when you talk about a drumbeat, that’s like a hint of something that may not be true. You might be mishearing something out in the forest, out in the jungle. But if you hear a drumbeat and it really is the beginning of the attack of your campground, that’s something real. But more likely it’s just some noise.

And so what I’ve heard, which I call a drumbeat, and I’ve yet to confirm it, is Warsh has said multiple things that are interesting to me. One is the balance sheet. He seems to be, along with some of the other members of the current FOMC, interested in shrinking the absolute size of the balance sheet and changing its composition. Neither of those things are suppressive of long-term bonds. They’re both removing suppression from long-term bonds. Now, they haven’t done it. They have a task force that’s studying it, and I’ve been advising the Fed a number of times over the last years and writing in my pieces that there’s an easy way to do this, but they just have done nothing about it. So again, just a drumbeat. That’s what he’s saying his direction is.

And then more recently, he’s been of two minds, but more recently at the July FOMC meeting, he said five or six times that he is happy that the bond market is doing the work of the Fed. What does that mean? He was referring to higher long-term interest rates were occurring, and he knows that that’s a tightening, and he likes that, and that’s new. And he also repeatedly said, “Let the markets figure this out.” And what that basically says is, instead of suppressing long-term interest rates, we’re just gonna take our hands off the brake. Which is a big change.

I don’t know if that’s the reason why long-term interest rates have been rising lately. Again, what’s interesting is long-term interest rates have been rising while every major piece of data has been on the cool side. Retail sales most recently, inflation, two months of benign to soft prints, NFPs that were weak, and yet the bond market’s selling off. So it’s possible that markets are front-running it. I don’t know.

But it would be a big change if Warsh decides to truly fight inflation by using balance sheet policy. Now, that said, just like Janet Yellen muted QT by financing the government mostly with bills, Bessent holds an ability to mute this effect as well. So he has to be on board. There’s no evidence that he’s on board yet. In fact, he’s done things that would say he’s not on board, like most recently with the yen intervention. He gaslighted that a program called FIMA was being used to avoid Japan having to sell bonds to finance the intervention. It wasn’t used, but he said it could be used and then later said, “Well, we plan on using it.” Those are things that would say, “Hmm, maybe Bessent just isn’t gonna let this happen.”

But so far, I’m now looking out. Last time I noticed this was in December of 2021, where I wrote Drum Beats of QT because Powell just hinted at a press conference that they’re looking at the balance sheet. Markets didn’t react, but when they announced in January that they plan on doing QT, they sure did, and they continued to react for another six, eight months.

So the next set of meetings are the interesting ones. Jackson Hole will be the next time. Warsh is more likely than not gonna continue to rest on the task force’s doing their work, one of which is the balance sheet, instead of making bold policy decisions. So I’m not expecting big news out of Jackson Hole, but yeah, I wanna read the tea leaves there. Then we still have the midterms, and so anything really disruptive to me is more likely to happen after the midterms than before. But I have my ears peeled for more drum beats.

Jack: Yeah, that was one of the things I liked about what you did in this piece — you were not saying, “This is over. It’s changing tomorrow.” You were just saying, “This is something kinda just slowly coming up behind the scenes, and I’m keeping my eye on it going forward.”

Andy: And I learned that because I got a little bit more short equities than I would’ve liked when I heard the drum beats in December of 2021. It took six weeks of pain before it ultimately paid off.

I just don’t think markets are... I mean, the basic point that I think is coming from the market consensus is there are some people that think inflation’s gonna just be okay — there’s a transitory group of people out there. That’s both in the Fed and private sector analysts. And then there’s a whole bunch of people who, for whatever reason, even after just two meetings, have already made up their view that Warsh is not a serious guy, he has no credibility. And that explains part of the market narrative. Certainly explains gold, explains bonds. They could be right. I don’t think they are, and I think they’ll be surprised, but that’s the narrative that’s playing through markets, and so I’m just not gonna run in front of that.

Jack: Do you think the way this is playing out makes them less likely to do anything with short-term rates? Like, I heard someone argue the other day that basically Warsh wants to focus on his task force and all his long-term stuff, and he doesn’t wanna distract from that with short-term rates, and long-term rates are coming up already. Do you think all this makes it less likely they do anything with short-term rates?

Andy: When he was confirmed, I said June, July, September, and October are all dead meetings. I still think that’s true. Why would he... I mean, there was always the possibility that there would be no cover, right? You get super hot inflation numbers in July and August instead of the opposite, the committee’s gonna wanna hike. Three of them already do. They only need four more to override him. He’s not gonna be forced to hike. But now he has cover. It’s working out for him.

So I think he can delay through year end, till December, through the midterms, till after his task forces have completed their work, and then we’ll see. And if inflation’s still robust, he’ll have to hike.

Jack: I wanna talk about the quarterly refunding announcement part of this. Can you first, before we get into that, just explain what that is? ‘Cause that’s something you brought to my attention that I didn’t really know a lot about before.

Andy: Sure. So people like me, and particularly government bond traders, have been following the quarterly refunding announcement for decades. And it basically says — the Treasury announces every quarter what its future auction schedule’s gonna look like, how many of each bond they’re gonna offer. And it had been a sleepy report until this big monetary shift where big movements of Treasury issuance had to flow through the market, big changes.

And so what it is, is there’s a certain amount of financing that the government needs, and they just tell you in this announcement how they plan on doing that. And so what I care about is, is the financing need growing, and is the composition of the issuance changing?

And frankly, it was really important in 2022 and 2023 to be all over that. Since 2024, the auction sizes have been fixed. They haven’t changed them at all for literally two and a half years. And so there’s been no signal from that. But every quarter you have to pay attention, because if they decide to increase the coupon issuance, that’s gonna be bearish assets. And if they decide to decrease and use bills more to finance, that’s bullish assets, just roughly. How extreme depends on the size.

But they’re running into a constraint, in that the debt just keeps growing. The deficit itself is going to grow this next year. And that’s gonna force them to increase auction sizes and have that tightening effect within the next year.

Jack: So this cannot go on forever, basically, is the idea, right?

Andy: Well, I mean, so there’s two things. There’s how the government is financed overall — bills versus duration proportion — and then there’s the overall size of the issuance. Like, the bigger the deficit, the more we’re gonna have to issue. That’s of everything, of bills, coupons, everything. That’s a big pressure already.

There’s a less clear pressure of, for most of the last decade, two decades, the government has financed itself with longer term debt and not T-bills, and now it’s up to 21 and a half percent T-bills. And not only is the percentage high, but the absolute size of the debt stock that needs to be rolled frequently, many times a year, has grown to a large absolute number.

And so the question is, are they ever going to term out the debt? Meaning shift 2%, 3%, 4% of the government financing from bills to coupons. Bessent said he wanted to do that, but he’s shown no interest in doing that. So that’s another important aspect: does the government decide to term out the debt to reduce the amount of bills? Ignoring the new debt, but just the old existing debt — use less bills and more coupons, or more bills and less coupons. They could choose either. If they wanna tank the dollar, protect the long end, stimulate demand, stimulate stocks, they could go to 30% bills. They can choose whatever they want. So it’s a very important lever, but right now, for the last two and a half years, that lever has been just nonexistent. It just hasn’t been pulled.

Jack: So if this drumbeat gets louder and louder and we start to see less suppression of rates, I would assume this is not a great picture for stocks and bonds. And I also think about, what does it mean for AI CapEx? So do you have thoughts around that?

Andy: Well, you can’t get everything you want in life. If you wanna kill inflation, you have to do certain things. If you don’t wanna kill inflation, you don’t have to do those things. So it’s all a matter of choices. If you wanna protect the long end, you continue to suppress interest rates, but that’s gonna be inflationary and dollar negative. It’s gonna make it easier for companies to finance their AI build. That could be a choice. I’m not saying that it’s obvious that they’re going to choose to finally kill inflation. They might not.

Jack: These concerns about the long end going out of control, you hear those all the time. Are those overstated by a lot of people, you think?

Andy: Yeah, they sure are. I think you have to think about who the natural buyers and sellers of Treasuries are and where the demand comes from. And I’ll tell you, if I saw twos-tens at 200 basis points — 10-year yields 200 basis points over two-year yields — I’d load the boat. That’s a point where, on a duration neutral, interest rate neutral position, you get to make 200 basis points of positive carry. You don’t even need much leverage with that.

So I think there’s just incredibly deep demand at a price for duration. The question is, are you comfortable at five and a half, 6% long-term interest rates with three and a half, 4% short-term interest rates? Are you comfortable that that’s okay? I think it’s very normal and would be easy to continue to roll and finance, and healthy for an economy, great for banks. A little expensive for mortgagees and for corporations issuing debt. But a disaster? I mean, I’d imagine a world in which they went up 500 basis points. The only way I see that is if inflation goes through the roof. And that’s not in my crystal ball at this stage.

Jack: So I wanna shift to your other DSR on issuance. And you brought me back to my childhood here with this idea of the hamburger thesis. So can you talk about that?

Andy: Sure. So last fall I noticed that there was a reasonably meaningful shift in the way CapEx was going to be funded. Obviously, we knew since early 2023 that CapEx was gonna go through the roof, and it’s surpassed expectations over those years. So now we have a pretty large expectation for CapEx — six, seven hundred billion this year, a trillion next year. Big numbers, primarily done by the hyperscalers, but also some of the neoclouds, and even the frontier models. They don’t need it for CapEx, but they’re experiencing cash burn.

So when you look at all the money that’s been promised — and when I say promised, every semiconductor stock depends on this CapEx occurring. Every hyperscaler depends on the compute that they build, and then the revenues that’s generated from that compute, to meet their earnings forecasts. And those earnings forecasts are for parabolic growth, 20, 30% type growth of the overall pie of the tech world. Well, in fact, the whole S&P is expected to grow in the mid-20s in terms of earnings growth. All that depends on that CapEx.

And so prior to basically a year ago, all the CapEx was being spent by the accumulated cash from all these companies that were essentially CapEx light, and had accrued all these earnings over the decades selling us services without much capital, without much infrastructure. So they had all this cash, and they also had operating cash flow that was sizable. And they could use that to spend on CapEx.

And basically a year ago, they finally ran out of that stuff and needed more than their combination of cash on hand and operating cash flow to fund their CapEx. That requires corporate bond and equity issuance.

And the first big one was Oracle. You might remember, with all this CapEx and all this excitement around data centers, Oracle announced a big data center investment. And the stock — I think it might have gone to 300. Mid 200s at least. It just skyrocketed. And then they told us how they’re gonna do that, and it was through an equity issuance and a debt issuance. And the stock halved.

And that’s what I was saying. So J. Wellington Wimpy, who’s a character in a cartoon called, I believe, Popeye the Sailor, which is before my time despite my gray hair — the joke with him was that he would be a beggar, begging for someone to loan him money for a hamburger today that he would repay on Tuesday. And of course, the implication was he was never gonna repay that.

And so the hamburger theme that I have is that that’s what the AI companies are telling us. We’re gonna build you a data center, a hamburger. You just loan us the money today and we’ll pay you back on Tuesday.

And the numbers have gotten really large. They’ve gone to a point where I plot the total shares outstanding provided by all issuance and all share repurchases over the decades. And for decades now, the US corporations have net retired their shares through share repurchases and lack of issuance. They just didn’t need the money. And it’s a big number. It’s a trillion dollars of net share reduction per year that these US companies have done. Which, if there are fewer shares and just as many people wanna own shares, the prices go up. And so that’s been part of the reason why stocks have done so well, is the passive bid from corporations retiring shares.

And that’s shifted. Next year, I expect it to be negative. Like, instead of a trillion reduction, we’re gonna be zero or possibly net issuance. While at the same time, corporate issuance has blown up in terms of total supply.

And so that’s basically the issuance story, which is we’re in a period of time in which massive financing is needed that is tapping the public markets. The money that they collect is going into the real economy, which is stimulating the hell out of the real economy. That’s a particularly bad combination for bonds, where you’ve got stimulative investment at the same time as massive issuance. That’s a double whammy for bonds. It’s not so bad for equities yet.

But then you look forward and you see what’s coming. Not only are the hyperscalers gonna have to start funding, but companies like OpenAI go cash flow negative because they’ve been buying all this compute to train their models and not making much in revenues — though there’s a lot of talk about their most recent sort of snapshot revenue and what it implies for future revenue. We’ll see. But that financing is coming. And so I think we’re in a period of time where you have significant headwinds for assets, which works for the Fed, coming back to the other thing, while stimulative investment is going.

Jack: It’s funny because we ask this question all the time. We ask people about the ultimate ROI from AI. And to be honest, on both sides, nobody has any idea. They’re just trying to figure it out. But you’ve argued in this piece that this funding thing is actually much more important than that question, right?

Andy: Short term, right? I mean, of course, if everyone thinks the ROI is gonna be high, they want in, and so that makes the funding easier. But I don’t think anyone’s gonna know that the ROIs are high today when they’re making their investment decisions. They’re gonna be suspicious. There are gonna be some optimists, there are gonna be some pessimists, there’s gonna be guidance. All those things are gonna play through, and you’re gonna get a market price.

But then the stuff actually has to be sold, and for a period of time, there’s gonna be a concession demanded by the investors to absorb all this. The real issue is if all this gets through the system — all the issuance, the trillion next year and ongoing issuance needed to fund all this — gets through the market, then we can start focusing on the ROI, then we can see how it plays out.

But my bigger concern, and I’m not certain this is gonna happen, but this is my concern, is that I’ve seen capital markets work for my whole career, and the way they work is everybody is incented to get these deals done. Lots of incentives work through the system. I brought up a Substack that I thought was mostly just funny but also true, that in this process everybody lies. The issuer, the investment banks, the buyers, all of them are lying, because they’re all incentivized to get these capital markets deals done.

And so for a period of time, the capital markets roar, and then investors look at their portfolios and say, “Jesus, what do I own?” And the capital markets close. They just close. It just happens. Happens every time. After a huge burst, the capital markets close.

So the question is, do they close before the trillion is through the door? If they close at $500 billion and there’s $500 billion of new funding that’s needed to fund the CapEx, what happens? If it lasts for six months, in the grand scheme of things, does it really matter? No. Six months later, the capital markets reopen and the funding gets done. But for the time being, this CapEx circular finance machine just, you know, pencils down. Nobody can do anything. And all the ROI stuff, people start saying, “Holy shit, what if the capital markets never reopen?” Of course they’re gonna, but at what price, et cetera.

So what I’m basically concerned about is that before the funding gets done, the camel’s back breaks for a period of time and the capital markets can’t absorb it. During that period of time, you typically get a fairly big market reaction, because even though the supply is gone, now people are like, “Holy crap, look at the future earnings I have baked in based on the supply getting done.”

So that’s what I’m concerned about. Do I think it’s happening tomorrow? No. I think the capital markets are pretty good. My guess is they’re gonna get this Anthropic deal done, then they’re gonna go with OpenAI and the ongoing issuance from everybody that needs money. And I don’t know when or if it’s gonna crack, but that’s what I’m watching.

Jack: Just one more from me before I hand it back to Justin. One of the reasons I love following you is whenever one of these financing deals comes, or anything that’s going on behind the scenes in Wall Street, I love going to your Twitter, ‘cause you can actually break it down. And this $500 billion deal that Nvidia had with Wall Street firms — I had no way of distinguishing myself, is this like some red flag that we all need to be worried about, or is this some run-of-the-mill thing? So I was wondering if maybe you could break down what you saw when you looked at that deal.

Andy: Sure. So it’s an interesting deal. Prior to this deal, when a client — a neocloud, say, for instance — wanted to buy some Nvidia chips because they need them for the data center, they have to have them, Nvidia helped them finance that purchase, often with an equity stake for that, with arranging a third party to lease some of the data center capacity once it’s built, for stopping out the old chips as they depreciate, meaning being willing to repurchase, essentially leasing the chips to the data center like you would lease a car.

And each of these obligations was an obligation to a single entity data center, but it was an off-balance sheet credit enhancement that was offered by Nvidia to these companies. And they announced this $500 billion thing, but at the same time, they announced a similar one with OpenAI, in which they would be a credit backstop for $240 billion of their chip investment.

So most of the time, Nvidia’s been doing these direct deals. This is the first time in which they’re doing an indirect deal. So what’s happening is the data center guy still wants to build a data center. In the past, he would go and buy his chips from Nvidia, buy all his other stuff that’s necessary for the infrastructure from them, and borrow money from various private credit entities, et cetera, to get the money he needed to do that. And there might be side deals that enhance that credit, all those things.

Now it’s all just done in one. The asset, the data center asset, is deposited in this trust. They haven’t given us details about what it’s like. I call it a CDO. The collateral’s dropped into this trust. The data center buys an equity interest. Nvidia provides a credit guarantee for the assets — in other words, a stop loss for the chips and the whole assets. Credit enhances the thing, and that way senior secured and super senior traditional corporate buyers are willing to step up and lend money to this structure. People like BlackRock, people like Apollo, they’re gonna buy the equity tranche and the mezzanine tranches, the high grade, the high yield tranches of these things for their private credit investors.

But what it does is it softens... Because this whole structure can have multiple data centers contributing collateral, multiple counterparties in all these things, Nvidia isn’t facing the same counterparty with its credit guarantee, and so it’s attractive to them.

But in the end, it’s the same thing, which is Nvidia’s promising future support without having to finance it, off balance sheet. And so that’s normal. Like, this isn’t ringing the bell saying this is Enron about to happen. But you are seeing that all of this issuance is creating a little bit of concern, a little bit of indigestion, and these new structures are ways of tapping more demand.

And so to me, the signal was pretty simple. Nvidia is out of capacity to do direct deals, direct circular deals, but still very, very in need of supporting its customer, and just had to come up with another way to do it. Which to me is a slight negative, but we’ll see. You know, Nvidia reports next week. I’m sure their earnings and their forecast will be through the roof.

Justin: We’ve talked about this before, Andy, but let’s just revisit how both the issuance and the CapEx kind of plays into your not enough pie framework.

Andy: Right. I think the not enough pie framework is the ROI conversation, which is, in the end, the GDP of the nation — and I know it’s global, and there’s that — but the growth of the economy is what every market participant gets to eat from. We all want growth and returns in everything we do. So we wanna earn more money as a laborer. We want to get more earnings as a corporation. And it all comes from the same place, which is the overall economy, which is measured by the GDP.

Within that GDP pie, there are winners and losers. Right now, the winners have been chip makers. They’ve been taking a growing and disproportionate slice of the pie. Well, if the pie isn’t growing, that means somebody else is getting less pie. So you have to think about who that would be and what does that mean for the potential future growth of the pie.

But let’s just focus on the pie growth. The pie grows because we get higher productivity or higher population. Well, the higher population isn’t gonna happen. Not globally with this global national populism sort of thing. Not with this president. We’re not gonna get any more immigration. So it’s all about productivity. So that grows the pie. The other thing that grows the pie is leveraging up, but leveraging only grows it for so long.

Anyway, you can have a pretty aggressive forecast for GDP growth. And then you say, “Okay, so the pie is gonna get really big.” And then you say, “How many claims for slices exist?” And what’s a claim of a slice? It’s Google’s earnings forecast. It’s consumption growth. It’s government spending. Everybody has a certain claim on the pie. With equities, it’s their earnings forecasts.

And when I add up the earnings forecasts of not only the major AI players, but the entire S&P 500 — assume most of it goes to the AI group — the amount of increase in that pie requires either a GDP that is so high over what anybody is possibly estimating that it could not possibly occur, or massive slice growth. Practically 50% more of the GDP that flows today to corporations has to flow to corporations in the future. It’s five, six, seven percent that has to go to corporations. They’re around 12 now. It has to go to 19 or 20.

But that has to come from somebody. And if it comes from the consumer, and it probably will, because they’re the ones threatened with job loss, and so they’re gonna have to dissave to maintain their consumption. Will they do it? Probably not.

And so when I look at that total, I think there are gonna be some massive winners — massive, massive winners, like historic winners. Nvidia’s already one of them, in my opinion. But when you add up everybody, there can’t possibly be, in total, winners. And that’s what I mean by not enough pie.

Justin: There was a recent Washington Post article talking about our national debt and, to your point, demographics as well, and this kind of statistic blew my mind. It was in 1952, there were six people between the ages of 25 and 64 for every person in our country that was 65 or older. So there were six people working for every senior. As of 2011, there were 2.7 people in that age bracket for every senior. And by 2030, there’ll be 2.5 people working. So that statistic really blew my mind when I sort of saw the chart and the long-term decline.

Andy: Right. And that’s the Japanese story. That’s the Japanese demographic that eventually will hit Europe first probably, and then the US without immigration. And that just tells you that potential pie growth is low.

Justin: What would you be paying attention to in terms of issuance and supply? Is there anything that you’d be watching that would sort of tell you that, okay, this is the point where the market—

Andy: Bell ringer.

Justin: Yeah, the bell ringer.

Andy: You know, we’ve seen this before where all of a sudden there are no SPACs. We go from lots of SPACs to no SPACs. Capital markets close for no particular reason and suddenly. Sometimes you actually get the market reacting after — like the capital markets being the reason why the market sells off, but it’s pretty rare. The more likely thing is the market sells off, and then people cancel their deals. And then the market sells off, and then people say, “No, I’m never doing a deal,” and then we’re closed.

The idea that we have a cathartic... It’s possible that we have a — I don’t know, they price OpenAI and that’s the bell. It’s possible. Pretty unlikely, given the interests involved, to have that be sort of the... You know, I look at the AOL Time Warner merger, that felt bell-like, but it still took two months before the market closed.

So I wish I had that, because then I’d position myself. But I think this is the type of thing that’s gonna surprise us, and it’s gonna be more severe. If it were to happen, it’s gonna be very severe. So it’s by nature gonna be a surprise.

So the objective — well, the way I’m handling it, I’ve framed that in my DSR about a bubble, which is available publicly. How to manage through a bubble-like regime, which is what I think we’re in, is: make sure you’re at your risk target. Make sure you’re not using extra leverage. And I have a number of other things which I think we’ve discussed on the show before. I think you just have to review that and make sure that you don’t find yourself way offsides. And I think a lot of people are.

Justin: Do any of the names, I’m thinking like SpaceX or... I mean, would you... I forget if you own individual stocks. I believe—

Andy: I don’t.

Justin: You don’t. Okay, so it’s all assets. I was gonna say, do any, would you ever consider any of those as they come to market? But I guess the answer’s no.

Andy: I play around in my personal account, which I call my degenerate account, Degen account. Trading for lunch money when I’m bored. But I think the SpaceX thing’s interesting. I mean, I think it’s incredibly overvalued and isn’t possibly going to meet its... It’s an Elon stock. But at the same time, the unlock is really going to bring a lot of supply to market, and so I think I actually have a degenerate short in that from, like, 145. Where is it now? It’s probably up a lot.

Justin: Right.

Andy: But yeah, I play that every once in a while, but I’m just a fucking degenerate when it comes to that stuff.

But to be honest, when I think about that, what I do for my macro is I’m tracking every single issuance that’s occurred in the last year — well, and frankly, lots before that, but particularly every issuance. So I really care that Google sold its stock at three fifty-five. I believe three fifty-five. It’s at three forty-two now. Huh, eighty billion dollars sold at three fifty-five, it’s trading at three forty-two. That’s underwater for all the people that bought that. That’s important to me. SpaceX did its deal at one thirty-five. It’s trading one forty-three, but it was two hundred and something. That’s important information.

All corporate spreads for all the issuance that have happened in the last year are underwater. All the bonds have wider spreads than they were issued at. I’m interested in that because that tells you how much concession it’s going to take to absorb, and it’s gonna create the tension between, do I really wanna tap the capital markets today? If not, I can’t do the CapEx. So am I forced to? If I’m forced to, I may have to experience a big concession. I really wanna understand that dynamic, and the best way to do it is with comps. How have other deals done? So yeah, I pay a lot of attention to it. Punting it? No, not really. Only for fun.

Justin: Yeah, no, that’s important though. It’s kind of the whole context of these issuance and what you’re seeing and where they’re priced and where they’re at, and that kind of gives you your overall view.

Andy: Right. That’s why I grade Treasury auctions. I’m famous for doing that on Twitter. Every Treasury auction, I give an A to A+ to D- grading, and I just do it because I wanna recognize what’s happening. Like, does it matter that Google’s trading below its issuance price? I think it does a little bit. Does it matter that Intel is now trading at its issuance price after rallying ten dollars off the issue? I think that matters. It matters because everybody who’s thinking about buying the next deal is gonna look at their portfolio and say, “Huh, I got this Intel. It was up ten bucks for a few weeks. It’s now unchanged. Do I really wanna buy this OpenAI? Maybe not.”

Justin: Look at the data, don’t be lied to. Maybe that’s the message.

Andy: Well, just pay attention, right? Pay attention. You’re never gonna know. This is unpredictable. You just wanna say, “How’s it going?” And I would say mostly, like I described in the DSR, the issuance is getting done. It’s getting accommodated. It’s not going smoothly. It’s not like there’s an intense need to buy this stuff by speculative investors like there was for SPACs for a period of time. But it’s going okay. But the size is mind-boggling. And that’s to come. That’s not what’s happened to date. That’s what’s ahead of us.

Justin: All right, Andy, good stuff. Thank you very much. People can go and check out your DSR report, can follow you on Twitter, and just keep up on your thoughts through those two, and Substack too.

Andy: Yeah, it’s an interesting time. It’s not happening tomorrow, but it’s just an interesting time.

Justin: Thanks, Andy.

Andy: Thanks, guys.