2026年8月随想杂谈

Today we’ve got a bonus Random Ramblings. I’ll explain why I’m doing it in a second, but I’m talking about 3 things that have been on my mind.

Number 1 is my favorite company, Strategy. Disclosure: I’m long a little bit of Strategy. They published a new 8-K, and I had to talk about how crazy it is that this company continues to trade at a premium, as well as the capital allocation there and how silly it is that I think it took them this long to develop it.

Then I’m going to tie that into Strategy, which had one great call on Bitcoin. I’m also going to tie it into the other thing I’ve been thinking a lot about: Situational Awareness’s blowup and the risk process behind it. Again, they had one great call, and then it kind of blew up.

There’s another thing I don’t even mention in the podcast: If you have one great call and make generational wealth on it, how do you judge someone as an investor on that? Are they gods because they had one great call? Is it just a coin-flipping monkey? I don’t know. I think they’re obviously very smart, with very questionable risk-management practices, but I don’t even mention that in the podcast. Maybe that’s a call for another time.

I’ll talk about the Situational Awareness blowup and related things. I’ve written 2 articles on the blog; you can find them in the show notes. Speaking of Situational Awareness, they had a great call that was a thematic trade: long AI, short AI losers. The tough thing about thematic trades is that, yes, you can call the theme, but historically there’s a price. People know what’s going to be a good theme for the most part, and the stock market is very good at pricing a rosy future with a rosy outlook.

Historically, that’s been a tough way to invest. But over the past 10 years, if you just said, “Hey, I think AI is going to be a winner. Let’s buy AI themes,” boom, you’re rich. You said, “Hey, I think GLP-1s are going to be a huge thing,” boom, you’re rich. Thematic trends have been very easy, but I find them difficult.

Relatedly, if you’ve got a great thematic winner, when do you know when to sell? If the price is 1 or 5,000, it doesn’t matter. You still think AI is going to be a winner. So how do you know when to sell a thematic winner? I’m just rambling on a little bit about all of that.

So, we’ll get there in one second. But first, a word from our sponsors. Today’s podcast is sponsored by truta.com. Look, if you’ve been listening to this podcast for the past year, you know what I’m going to say. If you like this podcast, you are going to love Truta. Truta is two buy-siders who hop onto a call and discuss stocks that they’re interested in. And I’ll give you one. I have been really interested in the power shells recently because, on the heels of the Situational Awareness blowup, Situational Awareness was really long power shells like Nebius and CoreWeave and several others, and the stocks were all slaughtered on the Situational blowup. So, I was trying to get up to speed on these names. It seems easy: Oh, they just lease data centers. But there’s the devil in all sorts of details. How many more megawatts do they have to lease? What are the terms of the leases? Can you trust the management teams? All sorts of stuff. And Truta has recent calls. I’ll link to one. There was a Galaxy call with a very handsome anonymous buy-sider on one of them. I’ll link to them. It’s just such a great way to get up to speed on any name you’re working on. And here’s the great thing: Now they’ve got an MCP that connects to Claude and ChatGPT and all these things. So, whenever you’re researching a name, the first thing I do if I’m looking at a name, I say /truta, have it run the Truta skill and pull what people are actually talking about, what buy-siders are really talking about, what will drive and move the needle for the stock. So, look, if you like this podcast, you’re going to love Truta. And if you’re trying to build out your AI skills and get up to speed quicker on names, guess what? Truta’s got you covered with a unique database that no one else has. So, go to truta.com, and if you want to see the Galaxy call that covers some of the stuff that I talked about—and by the way, disclosure, I’m long a little bit of Galaxy—go to truta.com/glxy and you can see a preview of that call. So, thanks, Truta, for sponsoring this episode, and now let’s get to the podcast.

Today is August 3rd, and I just did a Random Ramblings episode about 2 weeks ago, but I’m doing—I don’t know if I’m going to call it my August Random Ramblings or a bonus Random Ramblings. I’m doing it for 3 reasons.

Number 1, I’ve got 2 things on my mind that have just been going over and over and over again, and if I pour them out of my mind into a microphone, I’ll be able to move on. That’s the main reason.

Number 2, it’s raining outside, and I’m going stir-crazy. This is when I normally go on my afternoon walk and clear my head, and I can’t go on a walk because the weather is awful.

Number 3, I had a podcast guest cancel on me. So I had an extra hour, hour and a half, maybe more than that, because I prep for all my podcasts. I had a bunch of extra time that I’d budgeted and didn’t know what to do with. Between the rain driving me crazy and all this stuff—but you don’t care about that—let’s dive into the podcast.

I’m going to start off talking about the company I’m obsessed with: Strategy, formerly known as MicroStrategy. Disclosure: I’m short a little bit of Strategy and long a little bit of the preferreds. I just cannot get over this company, and I swear I’m going to stick the landing and pull this Strategy thing through this whole Random Ramblings.

They published an 8-K this morning. Let me see if I can pull up the 8-K while I’m talking. The 8-K says, “Hey, we sold $300 million of Strategy stock.” This is a massive company, so it’s not like this is a giant thing, but it’s $300 million of stock. They also sold $100 million of Bitcoin, took the proceeds, and purchased about $80 million of STRC preferreds.

What you have in Strategy now is a company that’s trading over NAV, diluting shareholders, and selling its Bitcoin in order to buy back preferreds at a slight discount on the open market and park a lot of cash on its balance sheet. You look at it and you’re like, “How is this company trading at a premium?” It’s just so crazy to me. I can’t get over it.

The reason this is going to be a throughline is that I read their earnings call, as someone with a position in them should read their earnings call. It’s crazy how they can talk like it’s magic. They say, “Hey, look, Strategy’s evolved. A year ago, 6 months ago, 2 years ago, when we first started buying Bitcoin, we were a one-way capital issuer.”

What that means is they only issued stock to buy Bitcoin. Now they say, “We’re a multi-way capital issuer. We’ll issue stock to buy that preferred at a discount, or we’ll sell our Bitcoin to buy stock at a discount.” They’re saying it like they’re some genius, and what that means is—I mean, all they’re talking about is good capital allocation. I mean, not even good, just normal capital allocation.

Now, say what you will about crypto. You can be a crypto bull, a crypto skeptic, whatever it is. I’ve generally been pretty skeptical of crypto, but MicroStrategy’s underlying observation back in 2020, when they made the Bitcoin pivot, was a good one. You can say that on a whole host of different levels.

I think Bitcoin was trading at about $20,000 when they made the pivot, and it’s at $65,000 today. It peaked at about $120,000. A triple in 6 years—more than a triple in 6 years—is a very good result.

Now, we’re going to ignore that they kind of piled in at the top, so they’re actually underwater on their overall purchases. But if you just take the first thing, a triple in 6 years is a very good result.

They were also right on their theme. Their theme was, “Hey, there are equity people who want exposure to crypto in some way, shape, or form, and they cannot get it.” Back in 2020, it’s very easy to forget, Bitcoin was hard to get. You could get hacked. There were no Bitcoin ETFs, all that sort of stuff.

A public company going and doing this, and being able to trade it in your brokerage account—I was skeptical. I said, “Hey, these are financial markets. People get it.” But they were betting that people wanted a 1-click button to buy Bitcoin, and they were right. So they hit it out of the park with that.

But they never sold, they never traded, and they never even considered it until recently, when they evolved. I think they evolved because of their capital structure. Personally, I think they completely bungled their capital structure. I think they would have been better off with the kind of HODL-forever thing, just based on the story they’re telling. They kind of failed, but now they’re selling.

I try not to call myself an investor. I try to think of myself as someone who researches, but there's a lot of research that suggests investors as a whole are very good at buying and very bad at selling, right?

If I'm using myself as an example, if I am researching a company, I will research it a lot and I'll say, “Hey, this stock's at 10, and I think this is a really attractive risk-adjusted opportunity, and I think it's worth 20,” right? Then I'll start putting into my head, “Hey, here's where I think I'll start selling.” If it's at 10 and I think the stock's worth 20, then I probably start taking a little bit off at 15, get heavier and heavier at taking it off at 17 or 18, and sell it all at 20.

Obviously, there are 1,000 things that go in between that, right? Maybe in between now and then I find something that's trading at 5 that I think is worth 20. Nothing on this podcast is tax advice, but if it's 2 weeks from now, the stock's at 14, my cost basis is 10, and in 2 weeks I go long-term, I'll probably consider the tax consequences. So there are a lot of things, but in general, I've kind of got the game plan when I'm coming in.

I don't think I'm alone in that, but I have been thinking, “Hey, when I've got this game plan I'm coming into and I'm saying, ‘Hey, I'm buying this at 10, and I want to start selling it around 15,’ am I mentally falling into that sell trap that I mentioned, where investors aren't as good at selling?” I'm not putting as much thought into the sale, right? That's kind of mechanical, in the same way that the strategy before was mechanical: We sell our stock and buy Bitcoin, and we never do anything else.

If I'm saying, “Hey, I'm going to wait till the stock hits 15 and start selling,” am I getting too mechanical? More importantly, am I failing to incorporate new information? Because I talked about how it's very difficult to execute—at least in my opinion, probably as somebody who's not a great trader, or who still needs to evolve and improve—but a lot of the big money is made when you buy a stock at 10, it goes to 20, and you say, “This is better today than it was yesterday. I need to buy a lot more.”

That can happen particularly with events, but it can happen with fundamental investment theses, too. A lot of the big money is made—well, if you are saying, “Hey, I bought the stock, I think it's worth 20, and I'll start selling at 15,” that big clearing event might happen, and it can be very difficult for you to flip your mind because you've said, “Hey, I'm anchored. I'm selling at this price.”

The stock goes from 10 to 15, you start selling, and then the good news happens and it goes to 20. It's very difficult not to sell more, or to stop selling, or to buy more because you started selling at 15. So I have no great answers to that, but it's something I've been thinking about.

Again, this Strategy thing is crazy, and it cracks me up every time. The company's valuation seems so clear to me: This is a completely busted capital structure that needs to be reset, and it needs to be refinanced. But just the fact that they said, “Hey, we're going to start thinking about selling. We're going to start thinking about evolving,” was hilarious to me. I can't believe they managed to pull it off, but it also got me thinking about selling.

The other thing that got me thinking about selling, and the other thing that's been on my mind a lot recently and has been on the mind of everyone in finance, is the Situational Awareness blow-up. I'm sure everyone knows about the Situational Awareness blow-up at this point. I've written 2 posts on the blow-up on the blog.

The other thing I've been thinking about, situationally along the lines of selling, is that they had a generational callout, which was basically “long AI winners, short AI losers.” They had that callout—I don't know, 2 years ago, whatever it is—and they went up like 10× on it, right?

They've got privates and all this sort of stuff that kind of saved them, because I think the fact that they had this massive Anthropic stake that's up a ton and that they couldn't really lever is actually why they didn't zero the fund out. But neither here nor there. They had this generational run—an unbelievable call, right?

The issue is, along the lines of not selling, they didn't know—or they didn't know they didn't know—I don't know what the reason is, but they didn't rebalance it after this generational run. Let's just make it very simple. They went long semiconductors and power, and short software. They did it levered up, but they did it when both were trading at 100. The semiconductors went to 400, and the software went to 20.

That's naturally a de-grossing transaction unless you keep putting more on, but they kept pressing and pressing and pressing and pressing it. Eventually, the software went from 20 to 30, and the semiconductors went from 400 to 370, and the whole fund blew up because they kept pressing it.

I was thinking, look, you've got this generational trade, and you didn't press sell. You didn't do risk management. Obviously, this is a failure of risk management; it's a failure of everything. But I was thinking about that sale, right? They had this great thematic call.

One of the issues with thematics is, when you have a thematic call—whether it's long AI, long GLP-1, or long, choose your great theme—there's no price, right? Historically, the way markets have worked has been that if you had this great theme that everyone knew, if it's in the news, it's in the price.

People would say, “Hey, we want to go long health care in 2006 because the population is aging, it's got all these demographic tailwinds, and it's a protective sector.” Well, that was kind of in the price, right? I think health care did pretty well, but these things are generally priced in.

The interesting thing about the past 10 years is that, with themes, generally you could buy them at almost any price. If in 2016 you said, “Hey, software is eating the world,” it basically didn't matter what price you paid for the next 8 years, until the 2022 growth reset and then ChatGPT came along.

I understand that's very much saying, “Hey, 8 years, and then the accident happened.” But if you said in 2024 you were going to go long AI, it basically didn't matter what price you paid. Same with GLP-1s.

I've just been thinking: With themes, the tough thing is, if you call a thematic winner right, how do you know when the price has caught up to you? I don't really invest thematically. I invest micro for the most part. I look at individual securities and weird setups. I try to look at setups.

I will do some macro when I'm trying to get a lot better at this, because I've seen it happen a ton, particularly recently. A great one is software, right? I kept saying, “I want to be long. I see panic in software. I want to be long software,” but it's too hard for me because these are not terminal zeros. There's a chance of all of these being terminal zeros.

In March and April, you saw companies across the board giving their executives really juicy compensation grants at software companies. That was kind of your buy signal, and the software index is up like 50% since those bottoms. You had this buy signal of insiders getting greedy, and that’s one way I’ve been trying to expand my micro to macro.

But with thematics again, if you were long AI winners, where's the top? When does it matter? That's just something I've been thinking about. Obviously, there's a failure of risk management, but you had this great AI trade that turned out right. How do you know?

Just a few more things: I can't believe his wedding was the same weekend he blew up. That's just a disaster, you know? It's easy to say, “Oh, go have fun at your wedding. You kind of solved the blow-up, and you can get back to work on Monday.”

If you've ever had a lot of stress at work, it is very difficult. You cannot turn your brain off, and I am sure he just had his whole fund implode out of nowhere inside of a week. I can't imagine going to that wedding, thinking about that, and having that hanging over you.

The other side of that I want to talk about is the opportunity in these names. I mentioned this again in a blog post, but it's just really interesting. A lot of these names have come down a lot, and you've heard a lot of investors over the past few years say—and these are generally investors who missed the AI trade, and you can throw me in, too, although I don't believe this line of conspiracy theory—that a lot of the people who've been the biggest beneficiaries of the AI trade are either public-to-private crossover investors, so they were investing in the privates, or they had really interesting ties to the AI trade.

This situation is a great example, where the guy is a former OpenAI researcher and is, I suppose, now married to the chief of staff at Anthropic. What you'll hear people say is, “Hey, all these guys over the past 18 months had these generational runs going long AI stocks, and it wasn't particular genius on their end. What it was is that they had MNPI—not MNPI in the ‘you will go to jail’ way, but just MNPI in the sense that they saw Anthropic's numbers and said, ‘Oh, it's pretty easy to see: If Anthropic's going crazy like this, we should go buy semiconductors.’”

Or, “Anthropic has to pay through the nose for memory, so we should go buy the memory stocks.”

You’ll hear that a lot. You can go back to the DeepSeek scare in January 2025. You would hear a lot of people say, “Hey, these guys bought the DeepSeek scare.” A lot of them say, “Oh, they saw OpenAI’s numbers and everything and knew there was nothing to fear.”

Or, in March, there was this big slowdown in semi prices, and you’d hear about all these guys piling in instead into the March sell-off, right before everything really ripped from April to June of this year. And you’d hear people say, “Hey, they had access to them.” I’m sure there’s a little bit of truth there, but to me, these guys had conviction, right? They were invested in all of this stuff because they had conviction in the AI trade.

And they had conviction, so yes, the numbers probably helped them build the conviction a little bit. It’s easy to stay long something that’s up a lot when you’re seeing demand exploding in front of you and maybe you’ve got those numbers, but I think they had real conviction here. But again, that’s where things get interesting, right? It’s very easy to say, “Hey, I’m long and I see memory demand through the roof. I see pricing through the roof.”

That’s a nice thing, but again, I would just say, if you look at these memory companies, there’s simply no way you can justify the current valuations, even if you take whatever supernormal profits you want for the next 2 or 3 years. There’s going to be a cycle at some point. These are commodity plays, and the back end is going to look really ugly once these get overbuilt. Unless you think, “Hey, we can just no longer ever catch up to memory demand.”

Anyway, that’s not where I was going. I was talking about opportunity. I think one of the interesting ones—the thing they were pretty much longest in—was the power build-out. The names that have been hit pretty hard are things associated with the power build-out: the former Bitcoin miners that have transitioned to AI data-center plays because Bitcoin mining takes a lot of power.

They had these shells. Bitcoin mining sucks as a business. It might be the worst business ever invented. It’s hard to think of a worse one, to be honest with you. But we can talk about that another time.

You have all these Bitcoin miners; they’ve got tons of power. You say, “Hey, Bitcoin mining sucks.” All of a sudden, all the AI plays need power, and they transition to AI plays. A lot of these things have been really hit. In part, that’s because Situational Awareness was long a ton of these things in big size. They were some of their biggest positions.

And because these things were a little smaller, they owned most of them. These things have been hit. They bounced a little bit since the Situational Awareness blowup, but they haven’t bounced as much as they—I don’t know if “should have” is the right word. They’re down a lot, let’s say.

And I think that’s interesting because, A, a lot of these you can DCF the contracts they have. A lot of these are not trading for much more than the DCF of the contracts they have. And B, if you were one of the people who believed, “Hey, all these crossover funds are trading on insider information,” all these crossover funds are telling you the fundamentals continue to explode.

Understand, that’s asking a barber if you need a haircut. They’re kind of talking their own book there, but they’re saying the fundamentals continue to fly higher. And the thing that a lot of them were heaviest into is also down right now. So, if you kind of believe that, there’s a really interesting opportunity there.

Do I believe that? I don’t know. But I do know contract securities, and a lot of these things—I mean, if you can buy them for the DCF of their contracts, whether it’s with CoreWeave or with Meta, and get everything else for free, that’s a very interesting call option to me.

I understand with CoreWeave, if you’re a real skeptic, you might worry about the credit risk there, but I think it’s really interesting. I’ve spent a lot of time doing work on them. I’m still doing due diligence, and it sounds easy, like, “Oh, yeah, they’ve got these contracts; you DCF them,” but how much of the CapEx is there? What are you putting on the terminal value? What are they responsible for? What are they not?

Some of these contracts have the customer responsible for the power, and in some of the contracts, the data center is responsible for the power. It varies a lot, and there are all sorts of different ones. Are they just doing a triple-net lease to the customer, or are they also going to buy the GPUs, which some of these have done? That’s a much higher-risk, higher-return business.

So anyway, I guess those are the things that have been on my mind. I’ve written about some of them. There are posts on the blog, but I had some extra time. It’s raining. I had no other way to get it out, and one of the nice things about running your own podcast is that, if there’s no way to walk around and sort through your thoughts, you can hop on the mic and just start rambling. So this is kind of the bonus random ramblings again. If you’ve got thoughts on any of it, feel free to reach out anytime, but I’ll talk to you guys later this month. 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.