2026年8月下旬随想

I'm on for my monthly random ramblings for the month of August. I did one earlier this month and had a lot on my mind, so I said at the time, “Hey, this will probably be the first of 2.” I figured I'd come on for my bonus random ramblings.

I'm going to talk about 4 topics today. First, we're going to talk about a post I published today on the blog yetanothervalueblog.com. What a great blog. You should go check it out. I published a post there talking about the equity risk premium and markets today. I wanted to elaborate on that a little bit. Then I want to talk about public CEOs using weird capital allocation, sticking with the CEO's trend, and how you would know if someone was the next great CEO.

So, we're going to get there in one second, but first two things. A, disclaimer remind everyone that nothing on this podcast is investing advice. That's always true, but you know, probably particularly true today because it is just me rambling around with things that are randomly popping through my mind. So, just remember I'm rambling. Full disclaimer at the end of the podcast and in the show notes. And the second thing I wanted to say is this quick shout out to our sponsor. This podcast is sponsored by Tratta. Tratta is two buy-siders talking to each other. Tratta records it, anonymizes, and publishes on the web. I think it is just such a fantastic fantastic way to get up to speed follow any new name because it's two buy-siders who really know what they're talking about cuz they're following the company they they follow it closely enough that they say, "Hey, I want to spend an hour of my time talking to someone else who follows it closely." So, I think it's just such an awesome product and if you follow this podcast, I think you're going to like it too. So, go check them out at tratta. That's t r a t a.com. And with that said, let's get to the rambling. The first thing I want to talk about is the equity risk premium: bond yields going up over time and how that affects equities. I published a post on this, so you should go read the post. Oh, I should mention it is Wednesday, August 26th as I am recording this. Very broadly, my post this morning was talking about how, in the mid-2010s, let's say Treasuries were trading at about 2% annualized. Stocks should have been trading at something like 25 times price-to-earnings.

The way you generally measure stocks is through an equity risk premium. This is the added risk, the added return, you get for investing in equities. There's lots of math involved in that. You have to estimate cash flows and all that sort of stuff, but you can kind of say, “Hey, if, you know, the outlook for equities is 29% and outlook for Treasuries is 25%, then the equity risk premium there is 4%, right?”

Historically, the equity risk premium has been about 4%—a little bit more, but about 4%. Treasury bills had traded at historically about 5% over the past 60 to 70 years. They traded from 5% to 2% in the 2010s. Just on that, equities should have traded down to, let's say, they stuck with a 4% equity risk premium. They should have traded for a 6% yield versus that 2%, right? Two plus 4 is 6.

That would have implied, based on growth rates and stuff—and again, lots of assumptions—that stocks should have traded for about 25 times price-to-earnings. But they did not. Stocks hovered in the mid-teens, which is kind of where they have historically. The way you explain that is the equity risk premium blew up and went from 4% to 6%. That's the 2010s.

Today, Treasury rates have been all over the news, especially for the past month, as Treasury rates have screamed higher. They're the highest they've been in 20 to 30 years. The 10-year is in the 4.5% range, if I remember correctly. The 30-year is even higher. The Treasury Secretary is intervening and buying the long end to kind of suppress rates.

The reason I've been thinking about that is I've been thinking, “Hey, what if you have a reversal of the 2010s right now?” Hindsight is 20/20, but from the early 2010s until today, stocks were on an absolute tear. Part of that is earnings growth has been very strong. Earnings are very important here, obviously, but another part is that they were starting with a kind of suppressed multiple versus interest rates.

As earnings growth picked up, the multiple also expanded, and you got this beautiful double whammy of strong earnings growth plus multiple expansion. Stocks did low-teens annualized returns for 10 to 12 years from the mid-2010s to today, and that's very, very good performance, obviously.

I've been thinking about that today. The equity risk premium is not that high. It's at about 4.5% right now, so it's kind of in line with averages. What's really happening is earnings are super strong. Maybe I should be worried about an earnings-growth wipeout recession and all that sort of stuff. Anyway, all of that is in the dark hole.

Here's what I want to elaborate on. There are 2 specific things. The first is the equity risk premium. I've been thinking: If historically bonds traded at about a 6% yield and the equity risk premium was about 4%, then equities would, based on the implied math, do 10%.

When Treasury bonds trade down to 2%, if the equity risk premium stays at 4%—and remember, in the 2010s it went up to 6%—then equities should do 6% annualized, right? I've been thinking, “Hey, that seems strange, right? Should the equity risk premium be compressing as Treasury rates go down?” Because 6% is 3 times 2%, whereas 10% in my example is 66% more than 6%.

I've been wondering: When interest rates are low, should you actually be expecting the equity risk premium to compress and go down? Should price-to-earnings multiples be expanding?

Why is that relevant? Well, if interest rates are rising today—and again, they're not out of line with historical averages—but if interest rates are going to keep rising, could you actually be looking at the reverse? Everything I've said is based on the historical average, but if you see interest rates continue to rise, should the equity risk premium also rise?

If historically it's been 4.5%—we're at 4.5% right now—should it go back to 6%, as it was in the 2010s? Because if you see interest rates rising and the equity risk premium rising, price-to-earnings multiples are going to fall a heck of a lot.

I've just been thinking about that. I'm not going to say that's foolproof, and this is more macro than I normally talk about, but I've been thinking a lot about it because interest rates have been on my mind.

The other reason interest rates have been on my mind is the earnings number I talked about. Earnings growth has been unbelievable for companies in general for the past 10 years. Some of that is the Trump tax cuts from Trump 1.0, taking the corporate rate down from, I think, 35% to about 21%.

That's fantastic for earnings growth, right? It's not just that the tax comes down. The big thing is the tax comes down, but you also get some added investment effects as things that are on the margin at a 35% tax rate become very profitable at a 21% tax rate.

You've had this strong earnings growth for the past 10 years, and for the past 3 years you've had really strong earnings growth driven by the AI trade. I think the AI trade is interesting, too, because a lot of it gets started in the last remnants of interest rates being low. They're not as low as they were in the 2010s, but a lot of these things are getting started and funded right before the big interest-rate spike of the past, let's call it, year starts.

I've mentioned on this podcast a lot—the power shells, the data centers. A lot of these are former Bitcoin miners that flipped to data centers, right? If you follow these companies, the way they generally do leases when they lease out to a hyperscaler, NVIDIA, AMD, CoreWeave, or whoever they're leasing to, is a 15-year lease with 2 options for the tenant at the end of the lease—5-year options.

If you do the math on the lease payments that CoreWeave, a hyperscaler, or whoever is paying, what it kind of covers is this: If the plant is going to cost $5 billion to build, the net present value of the lease kind of covers the $5 billion plus a little bit of risk capital for the developer fee.

What the company is generally betting on—I mean, they will make a profit on the projects even at the end of 15 years if there's nothing left in that data center, right? They will make a profit on it, but it won't be huge. The real thing the companies developing these are betting on is the terminal value of the data center they've built.

They're betting that in 15 or 25 years, if all those options are picked up, they can re-lease it. That re-lease is, quote-unquote, free for them because the build cost, on a net-present-value basis, was covered by the hyperscaler.

Hopefully you're bearing with me. I understand I just did a podcast saying, “Hey, the story is the whole thing,” and now I'm diving into numbers and DCF, but whatever. The reason I've been thinking about that is interest rates have been going up. Let's just say they went from 3% to 5% over the past 18 months. You can correct me on that—maybe 4% to 5%.

It does 2 things to the AI buildout for these data centers, right? These data centers are big, big projects. We can talk about memory and chips and all that, but a lot of the money is going to the data centers and the power. If interest rates go from 4% to 5%, it does 2 things.

Number 1, the lease payments that the company is making have to go up, right? A lease payment 15 years from now is worth one thing if interest rates are 4%, but it’s worth less if interest rates go to 5%. To counteract that, you need to jack your lease up by somewhere between 5% and 10% in the lease-rate numbers to counteract the effect of interest rates.

That’s number 1. Number 2, I just mentioned that the reason all these companies are building the data centers is that they get that terminal value for free, quote-unquote. If you said, “Hey, you build something, it’s going to work for 15 years, I’ll cover your cost of capital for those 15 years, and at the end you’re left with the asset, and that asset’s worth nothing,” you’re not going to do that, right? What person is going to do that?

You need to have a view on the terminal value of these assets and that they’re going to be worth something, or else you’re not going to do it, or you’re going to demand more in lease payments, right? If interest rates go from 4% to 5%, the terminal value of that asset goes down—or, sorry, the terminal value could be the same, but the NPV of that terminal value goes down, right? Again, because it’s 15 or 25 years out and you’re discounting it back to today.

The other interesting thing is: go look at what a Bitcoin miner was charging. I believe the Bitcoin miners cuz some of these Bitcoin miners did hosting deals with Bitcoin players. A lot of these data centers have come from Bitcoin miners, because Bitcoin miners need a lot of power for compute. They were the only people who needed tons of power for compute before the AI centers really started ramping up.

If you look at what they were charging on a dollar-per-kilowatt-hour basis, or whatever you want to call it, versus what they’re charging today to these AI players, prices are up about 10×, right? The other reason I think about that is: you’re betting on that terminal value, but you’re taking a lot of the risk that the AI buildout has blown up in the next 15 years, right? You’re going to get hit in 2 ways if you’re one of these data center players.

If the AI data center race is still going and all these data centers are still really in demand, almost all of them are contracting out with options for their tenants, right? The tenant is going to be very in the money on that option, and they’re going to pick it up. If it goes the reverse and the AI trade blows up, you’re going to get hit in 2 ways.

First, your AI tenant that gave you this 15-year lease might go bankrupt, right? None of these companies were here 15 years ago. You might have built this big project thinking you had a great tenant for 15 years, and in year 6, they’re bankrupt, right? That’s number 1.

The other way it hits you is when you’re releasing in 15 years, because now your tenant is not in the money on the option. Whenever I value these—or anyone values them—a lot of times I say, “Okay, we have a contract with AMD. It’s for 15 years. It’s for $100 million in NOI per year.”

In 15 years, we inflate it up from $100 million to $120 million, $150 million, $200 million, whatever it is, and we’ll put a 10× multiple on that and then discount it back to today, right? That’s what they use to calculate the terminal value.

My worry is that if the AI bubble bursts—and I use “bubble” in quotes. I don’t think it’s a bubble, but obviously it gets pretty frothy sometimes—if the AI bubble bursts, that lease that went for $100 million to AMD or whatever, if it had to go to a Bitcoin miner or the next-best player, going back to 2023, it would have been going for $10 million per year, right?

You could get hit doubly, where that releasing risk really hits you. Why does that matter? I’m tying this into interest rates because if interest rates keep creeping higher and the data centers are a material portion of the AI buildout, the data centers are going to start crowding out some of this AI investment, right? Could it slow down just because interest rates go up? It’s going to start crowding out investment.

If AI slows down, all of a sudden the data centers are going to start looking around and saying, “Oh, we’ve got customer credit risk. That terminal-value risk is looking a lot less attractive than we thought it was.” I don’t think we’re there yet, but if you took interest rates—just theoretically—to 100%, obviously we’d be there, right?

We’re not there yet, but I have been wondering: this is Finance 101, right? As interest rates go up, investment gets crowded out. Investments that you would have made when interest rates were at 2% don’t make sense at 5%. I’ve been wondering if rising interest rates start to have some impact on the AI squeeze.

I don’t think it’s going to get there if interest rates tick up another 20 basis points, but at some point it would have an effect. I have been thinking about the cyclicality and circularity of this. It’s very interesting to think about.

I’ve been rambling. I’ve gotten really wonky. Let me go to the 2 other things I wanted to talk about.

I did a post earlier this month, and there’s been lots of coverage of this on UWMC, that is, United Wholesale Mortgage. Their founder owns the Phoenix Suns. The company is a wholesale mortgage lender, and they got into this crazy bidding war.

In Q2, they came out and said, “Hey, we took this massive bath on a hedging loss.” They took this massive bath on a hedging loss because of a deal that they had lost. They were not under contract at any point in Q2, but they claimed they kept the hedges on and took a massive bath. They had to do a distressed raise—all this crazy stuff, right?

I’ve been thinking about that because if you listen to short sellers—and Hunterbrook has done a lot of coverage on this, and I think they’ve done very good coverage—UWMC was running its capital allocation to pay out big dividends. They were doing it not because they thought the dividends were sustainable, but because the CEO owned the Phoenix Suns and needed the dividend payments to pay for the Phoenix Suns, right?

I don’t know if that’s true or not. Those are the allegations. I think some of them make sense, but I also think that when you’ve got somebody who takes a massive hedging loss in Q2 on an asset that they weren’t under contract for, there might also be a little bit of cowboy in them, right? Who knows?

Go back to Cogent, CCOI. I’ve done several podcasts on Cogent. The company was paying a huge dividend even as its leverage was really starting to tick up as it integrated the Sprint deal. You can go find the podcast with Aaron Chen, who I think has done a really nice job covering the company, though obviously it has not worked out well.

They paid a huge dividend long past the point when paying a huge dividend made sense. I think there’s an obvious explanation why. The CEO said it on their calls: he owned a huge Washington, D.C., real estate portfolio, and he was getting margin-called over there.

I think he wanted Cogent to continue to pay big dividends so that he could fund tax payments on his Cogent stock and his RSUs as they vested, and so that he had cash coming off the Cogent stock to cover his D.C. portfolio. I’ve just been thinking about what happens when public CEOs run capital allocation that is designed more for their benefit and their personal balance sheet than for public shareholders.

It is extremely rare. The obvious answer is bad, right? Anytime someone does this, it’s classic management misalignment: management is doing something for the company that benefits them versus shareholders.

But it is quite rare, and I’ve been thinking about better ways to tell. How can you identify it? Because David Shafer was at Cogent was very clear, I'm having trouble with my Washington D.C. portfolio. You know, they disclosed the margin loans. He was very clear about about it, and it came back. Go pull up the Cogent stock chart. Now, there were other issues with Cogent, but go pull up the Cogent stock chart. Doesn't look great. UWM C never said, "Hey, we're paying dividends to fund the CEO's uh NDA purchase." But, it it it seems like what they did in hindsight. But, I have just been thinking like how could you evaluate this? Because I just gave you two examples, and these things are blow-ups, right? It's a disaster. How could you find companies that are being run for the management team’s balance sheet or the management team’s needs versus what makes the most sense for the company?

I don’t know, because companies do crazy things with capital allocation all the time. But there are 2 really clear examples, and I can think of a few more, loosely. I don’t want to rattle them off the top of my head and claim something that’s not true.

But it’s a pretty big red flag, and I’ve been thinking about that. And look, this is one of the reasons activists exist, right? “Hey, you’re running a capital allocation strategy that’s more designed for you than shareholders. Let’s go. Let’s get involved. Let’s change it.”

Now, it’s not lost on me that Cogent and UWM, you know, there is something to the fact that they were founded by their leaders. David is the founder of Cogent. Ishbia is the founder of UWM. There is something to, “Hey, you have the founders who, whether they control the company or not, are going to carry a lot more influence with it.”

You have the founders doing something that benefits them personally. And, you know, maybe there’s also an element of the founders looking at the company and saying, “This is my company. This is my baby. I can run it the way I want to.” But something I’ve been thinking about a little bit recently—and I’m going to keep an eye out for more places where management teams are kind of running the balance sheet for themselves.

Speaking of great management teams, the last thing that I’ve been thinking about is based on a loose text conversation. I think my friend who I was texting with listens to some of these, but even if he does, it was such a loose conversation that I don’t even know if he’d remember I was having it with him.

Let’s say this is the text thread. I’ll give you the background. My friend was talking about a company, and he was saying, “Hey, everyone in this company 3 years ago, 5 years ago, thought the CEO was the next Mark Leonard.” Mark Leonard is the founder and CEO over at Constellation Software, which is one of the best, if not the best-performing, stocks of the past 20 to 30 years, right? The Canadian software company.

Anyway, with this specific company, 3 or 5 years ago, everyone was talking about the CEO like he was the next Mark Leonard. He was going to build the next Constellation Software. And my friend said, “Hey, I knew the CEO pretty well. Good guy, really sharp. Anyone who thinks he’s the next Mark Leonard is completely crazy.”

In fact, he would say, “I would laugh at everybody any time anyone said, ‘This is the next Mark Leonard,’ right?” He’s not saying the guy wasn’t a fraud or a dope or anything. He was just saying, “Hey, this guy is somewhere between a B and an A-minus, and, you know, Mark Leonard would be an A-plus.”

That’s kind of where he was going. And I was thinking to myself, how would you know if someone was the next Mark Leonard, right? Because the thing with Mark Leonard is that he compounds this business over 20 to 30 years.

A lot of times, the person who is getting compared to Mark Leonard, and a lot of the wannabes, do start off with a great track record, right? What it is is they’re taking on hidden, enormous risk, right? In much the same way that every time someone puts somebody in a magazine and says, “Hey, this is the next Buffett,” based on the first 10 years of their career, I mean, the fantastic thing about Buffett is that he did it for 50 to 60 years, right?

A lot of times, when somebody says, “Hey, this is the next Buffett,” based on the first 10 years of their career, the next 10 years don’t look as good. A lot of times, you can get great results in investing by assuming a risk that you don’t even realize you’re assuming, or intentionally assuming a big risk, right?

You can get great returns in 1 year by levering up and YOLOing something. You’ll get great returns, but eventually it will blow up if you keep levering and levering and levering. You know, I’m looking at you, Situational Awareness.

How would you be able to tell early in someone’s career if they’re the next Mark Leonard? Again, with Buffett, at least you’d have the investment-returns track record. But with Leonard, it’s hard even over 10 years. The stock price could have started really low and gotten really inflated. These things are really volatile.

So I’ve been thinking—I mean, one of the common things I talk about on this podcast is, how do you judge management? How do you form relationships with management? You know, management teams are great salespeople, and as investors, I find most people—and I include myself in this—can get sucked in by a management team.

“Hey, why was this quarter bad?” And the manager says, “Hey, this was a one-time thing. It’s all part of the plan, you know.” And you say, “Oh, okay. I talked to the management. I looked them in the eyes.” And the management teams are always just taking us investors for suckers.

You know, I find what happens is you believe them until, after a year or so, you say, “Oh, these guys are liars.” And anybody new comes in, you say, “Those guys, you just can’t trust them.”

So, anyway, I don’t know where I’m going with that. These are my random ramblings, but I’ve just been thinking about how you separate out the track record for someone who’s great. Again, my friend was just talking about someone who was good, and everyone else thought they were great. My friend just kept thinking, “They’re good, they’re good, they’re good.”

I guess the other thing I would think of is, you know, a lot of the managers that people become enamored with, in much the same way that a lot of the wannabe Buffetts just have 1 strategy or 1 risk they’re betting against and make a lot of money, are actually managers who just found 1 great theme and rode it, right?

I think one of the interesting things about Mark Brad Jacobs, who is kind of struggling over at QXL right now—the building-products roll-up that he’s doing—is that he did multiple roll-ups in different industries and turned them into great successes. It’s very rare to find somebody who does something multiple times in different industries and has big successes.

I’ve been thinking, even with a Mark Leonard at Constellation, he bet on vertical software, right? That was a great bet, and it’s done fantastically. But that was a great bet. How much do you say the Mark Leonard genius is Mark Leonard versus, “Hey, he had a good industry insight”? If he had chosen coal, it wouldn’t have been as great.

There were lots of guys who were brilliant geniuses at oil when oil was going up from 2000 to 2008, or from 2010 to 2013 or 2014. And then you never hear from them again because it turns out, hey, they weren’t operational geniuses. They weren’t capital-allocation geniuses. What they were was people who had 1 bet on 1 factor, and once that factor stopped, it ended.

Constellation, you know, what would have happened if the AI trade—the SaaS apocalypse—had happened in 2014 instead of 2026? Would Constellation Software have been the same? I mean, we’re talking at the end of August 2026. A lot of the SaaS companies are bouncing back, have bounced back, and everything.

I think the huge fears of April—and I had some of these at the time—the huge fears of, "Hey, everybody's going to hire one software engineer and buy code all their internal software," I think those have moved to the side, though. There are some real terminal-value questions there that I’ve still got a lot of questions on.

But for a Mark Leonard, if you have that SaaS-apocalypse fear 10 or 15 years ago, does history look a lot different? It’s just interesting how the dice roll of the world and the environment you’re in can impact everything.

So, yeah, I think we’ve covered everything there. Those are the 4 things I kind of wanted to cover. Again, I’m just rambling. One of the great things about the post I put up about equity risk premiums earlier today or anytime I ramble on about random things, listeners reach out and they tell me what they think and sometimes I just say hey, thanks for that and sometimes they their emails really thoughtful and we'll have days, months, years long conversations on the topic. So, if anything in here struck a chord or you're interested in it, you know, always feel free, reach out, shoot me an email, shoot me a DM, and we can we can chat a little bit about it. So, that is my second random ramblings for the month of August 2026. I got some great podcasts coming up. My buddy Yaron Brook, I believe is coming on on Friday. One of the most popular guests hasn't been on in a long time. I think people are really going to look forward to that. So, we'll I'll I'll see you for the Yaron podcast in the near future, and we'll go from there. Talk to you soon. A quick disclaimer. Nothing on this podcast should be considered investment advice. Guests or the host may have positions in any of the stocks mentioned during this podcast. Please do your own work and consult a financial advisor. Thanks.