Aschenbrenner’s Margin Call: Why Leverage Upset the AI Thesis

Source: Startup Insider – Original Episode (episode in German)

Leopold Aschenbrenner, 25, a former OpenAI employee, built a hedge fund based on his 165-page essay “Situational Awareness,” which most recently managed $45 billion—long on AI infrastructure, short on traditional software, with a team of twelve people and investors including the Collison brothers, Nat Friedman, and Daniel Gross. In July, market dynamics shifted: The SpaceX IPO sucked liquidity out of the market, the anchor position SK Hynix fell by nearly 50 percent, banks issued margin calls—and Ken Griffin’s Citadel took over the publicly traded portfolio for $16 billion at a discount of about 10 percent. Björn Rieckhoff and Jan Thomas explain why the headline “35 billion lost” is misleading: Of the 45 billion in AUM, only about 11 billion was investors’ capital—the rest was bank-financed leverage. The decline in AUM doesn’t constitute a loss. And the episode’s central thesis: Aschenbrenner wasn’t wrong in substance; he was simply caught up in market dynamics at the wrong time—it was the leverage that toppled the portfolio, not the thesis. So you’ve got the valuation asymmetry between public and private markets, and what German fund managers should learn from this regarding DPI and divestment strategies.

Key themes in this episode

Transcript

This transcript has been edited for readability. The content and statements have not been changed. The original conversation was in German; this is an English translation.

Jan Thomas: Björn, hi, nice to see you! I hope you’re doing well. I was just thinking about you while preparing for this, since you moved away from Berlin a while back. I think you’ve told the listeners before that you live about an hour outside the city. So I was wondering: How must that feel for someone like you when you’re thinking about hedge funds? That’s such a stressful situation—in the world you’re currently navigating, you really don’t want any of that, do you?

Björn Rieckhoff: Oh Jan, of course you’re putting me in this kind of corner here, as if I lived on a farm in Brandenburg. But you’re right—I’m in the Mecklenburg Lake District. Exactly—it’s even more beautiful. It’s very, very beautiful. There’s infrastructure here too, there are smart people here, and above all, there are very stable—and in some cases better—internet connections than in Berlin. But sure, I also want to get a little away from Wall Street; you have to admit that.

Jan Thomas: For me, it was mostly about the stress, you know. I’ve seen pictures—you’re in a setting where you don’t want to be staring at your phone the whole time, worrying whether your portfolio is about to melt away, right?

Björn Rieckhoff: Nah, not at all. That’s true. You’re more likely to be thinking about which lake you’re going to next. That’s the luxury I have—that everyone here has. But yeah, that’s obviously a slightly different way of framing the problem—you’re right!

Jan Thomas: So maybe you should send Leopold Aschenbrenner a few photos and tell him: Life can look different, too. And with that, we’re diving right into the topic, I’d almost say.

Björn Rieckhoff: Let’s dive right into the topic. He’s only 25, after all. That’s pretty amazing, and it makes me think: If he’s living in the Seenplatte in ten years and looks back on his life—well, that’s definitely high-frequency, you have to say.

Jan Thomas: And he probably knows the area pretty well, since he’s actually from Berlin. And I think he was even a member of the Green Party. So he’s probably seen a bit of the surrounding area as well.

Björn Rieckhoff: Supposedly. I don’t know for sure myself, but I do know that he’s from Berlin—I think he went to the JFK School. I just looked that up while preparing for the podcast. He’s extremely young, in any case, and has been the subject of frequent discussion in recent weeks.

---

Jan Thomas: I have to say, I didn’t know him before. I mean, I didn’t just learn about him because of the crash—which we’ll talk about in a moment—but I’ve probably only really had him on my radar for the past four to six weeks. You’ve known him a little longer—or rather, you were familiar with his name.

Björn Rieckhoff: Exactly, I was familiar with the name, mainly because of an essay he had published. It was called “Situational Awareness” and was written in the early days of the AI revolution. He gave readers insight into how he, as an insider—he was employed on the AI safety team at OpenAI—viewed these developments. And I think he wasn’t the only one to draw this conclusion, but in this very, very long essay—165 pages, I believe—he developed an investment thesis in which he essentially established this “picks-and-shovels” model that we’ve also discussed frequently. Namely, the thesis that you don’t necessarily have to invest in model providers like OpenAI or Anthropic right now. It’s actually more about investing in the infrastructure that forms the foundation for building AI—that is, memory chips, data centers, power generation, and the like. So you’re not investing in the gold itself, but in the shovels used to dig up the gold.

Jan Thomas: Yeah, that totally makes sense—we hear about this “picks and shovels” thing all the time. And you just mentioned that he’s an insider, which is generally interesting. He used to work at OpenAI—he’s no longer there—but his girlfriend, or I believe his wife now, is also very well-known in the scene.

Björn Rieckhoff: To be honest—there have been quite a few tabloid articles about this lately: At the time his fund went under, he was at his wedding and stuff like that. To be honest, I haven’t looked into his family life.

Jan Thomas: His wife’s name is Avital Balwit, I think, and she’s Chief of Staff at Anthropic. That’s just—I think he’s really managed to build a network there, especially coming from Berlin and probably not having been in the U.S. for very long. Just look at who’s invested in his company. I think that’s really extraordinary.

Björn Rieckhoff: I’m just wondering—I know a lot of startup Chief of Staffs, but what exactly do you do as Chief of Staff at Anthropic? That’s probably the coolest role you could possibly have. On the one hand, you’re probably a jack-of-all-trades, and you have a lot of responsibility. But what responsibilities do you actually have?

Jan Thomas: And she’s actually the Chief of Staff for—there are different roles depending on how big a startup is—Dario Amodei, the CEO’s Chief of Staff. It’s really a perfect fit right now. But maybe also for insider information.

Björn Rieckhoff: Interesting, yes. But I don’t even know if she was already part of Leopold Aschenbrenner’s life when he left OpenAI. He was supposedly fired because of an information leak in 2024. He denies that, though, it must be said. And then he basically wrote this essay. That was, so to speak, the birth of his hedge fund, which we’re talking about today.

Jan Thomas: Exactly, that’s the reason we’re going into such detail in the first place. And a lot of what we’re saying now is, of course, an “outside-in” perspective—we’re looking at reports circulating in the market. When you say he denies it: We just don’t know—it’s mostly just rumors. Even whether there’s a Chinese wall in the bedroom between him and his wife and stuff like that (laughs). Who knows.

Björn Rieckhoff: (laughs) ... that’s actually the best way you’ve put it in a long time.

---

Jan Thomas: So, with that, I’d say let’s dive right into today’s topic. After all, a lot has happened.

Björn Rieckhoff: Exactly, a lot has happened there. Let me back up a bit. As I said, he built the hedge fund based on his essay “Situational Awareness,” in which he described the AI revolution. By early July, I believe he was managing $45 billion—with eleven other people, so a very small team. That’s a huge amount of AUM. Among those who invested were the Collison brothers from Stripe, Nat Friedman, Daniel Gross—a whole host of big names from the finance and tech scenes. And he himself essentially pursued an investment thesis built on two pillars. First, he went long—that is, he bought stocks in companies with business models driven by AI—memory chips, SanDisk, Micron, and the like; data centers such as CoreWeave; and power generation, including a few energy providers. As a result, he had a portfolio in which the five largest positions accounted for more than three-quarters of the portfolio. I’m mentioning this now because it’s relevant to the story that follows. The second bet: He bet against traditional software companies, meaning he went short on firms like Adobe—based on the theory that AI will undermine their business. On paper, this is a hedge; you have two different pillars on which to build your portfolio. But if you dig deeper, it’s the same bet, just from two angles: AI infrastructure up, legacy software down—one and the same market trend. And that trend reversed in July. I think for a few different reasons. In the background, you had the SpaceX IPO in June—the largest of all time. Anyone who wanted to subscribe to it needed money and sold something else to get it. Then you saw a wave of selling in stocks that had performed very, very well in recent quarters—and those included chip stocks as well. So you saw a dynamic in which this successful IPO siphoned money out of the rest of the market, on a scale that no one had really anticipated. Add to that the IPO stories of the major LLM providers, for which certain people are preparing—perhaps setting aside a bit of cash and selling stocks that have performed well. That put pressure on their portfolios. And then there was the listing of a position by Aschenbrenner—he had committed in advance to be an anchor investor there for around $7 billion: SK Hynix. The stock fell by nearly 50 percent from its June high, and that got him into trouble because he was extremely heavily leveraged in his portfolio. Of the 45 billion in AUM, only—in quotes—11 billion was his investors’ money—the rest was capital he’d obtained from banks. I’d like to explain this in a way that the average person—that is, the listener—can understand: If you buy an apartment for one million, put in 250,000 as equity, and borrow 750,000 from the bank, and the apartment increases in value by 20 percent, you’ve made 200,000. Then the apartment is worth 1.2 million, and you still only have to pay back 750. So you’ve almost doubled your investment, even though the apartment’s value hasn’t doubled. This leverage is real. However, it works just as well the other way around: If the value of the apartment drops by 25 percent, your money is completely gone. Then the apartment is only worth 750,000—exactly the value the bank wants for its loan. And then, unfortunately, you have nothing left. Exactly the same thing happens with a hedge fund’s portfolio. You then find yourself in a situation where the banks that gave you the money for this leverage suddenly call and say: “There’s some uncertainty here now; we’d like our money back.” And at some point, the only way you can do that is by liquidating the liquid stocks in your portfolio. Of course, this happens precisely at the moment when the value of that portfolio is no longer at its peak—one stock has plummeted by nearly 50 percent, and you still have to sell. That’s why all these articles are coming out now. In the end, it was Ken Griffin of Citadel who acquired the fund’s entire publicly traded portfolio for $16 billion, at a discount of about 10 percent. He essentially had to offload a portfolio of healthy companies to meet the banks’ margin calls—and lost money in the process.

Jan Thomas: We’ve seen this mentioned a few times now: Ken Griffin—I didn’t know him before either, though I’d heard of Citadel—is just incredibly experienced. People have been speculating all along that Aschenbrenner simply got an expensive lesson from one of the most experienced hedge fund managers. Maybe that’s part of the game. At the same time, take the example of apartments: With an apartment, the lease is fixed for a certain period. I can’t even imagine the scale of the phone calls Aschenbrenner is getting right now. When margin calls suddenly start coming in nonstop and the banks get nervous—that phone is ringing off the hook.

Björn Rieckhoff: Yeah, of course they’re extremely nervous—which is understandable, especially given the scale of it all. And to be honest: I have to be a little careful here, because I really don’t like it when people come down hard on Leopold Aschenbrenner, in a condescending way, as the German press in particular has done. After all, he has built up a great deal of value.

---

Jan Thomas: If I may jump in here for a moment, Björn: I think the early investors all cashed out with returns of 70, 80 percent or more. It’s more the later investors who are coming up empty-handed right now. That’s why I think his basic thesis wasn’t entirely wrong in certain respects.

Björn Rieckhoff: Not at all. And I think that assessment is still relevant and accurate. Two things I’d like to say about that: Everywhere you look, it says he lost $35 billion. That’s not wrong—but that’s the decline in assets under management, not a reported loss. He leveraged the capital threefold. So the investor capital deployed was never $45 billion, but $11 billion. That means he incurred a loss of about 67 percent in July. That’s still catastrophic, but it’s something completely different from saying “the $35 billion is gone.”

Jan Thomas: Exactly—and these days, as we know from venture capital, you often see paper gains. That’s really no different here.

Björn Rieckhoff: Yes, and that’s actually the interesting point—I’m glad you brought it up, because that’s exactly the conclusion I would have wanted to draw at the end of the analysis.

Jan Thomas: Sorry, I didn’t mean to jump the gun.

Björn Rieckhoff: No, no, it’s great—that’s exactly the analogy. What value remains with Aschenbrenner? The portfolio in the private markets—those illiquid investments that no one values on a day-to-day basis. What Ken Griffin has now taken over are the publicly traded market values that are priced in day after day. He still has his private investments in Anthropic and others—those weren’t financed with debt. They don’t have a daily market price. And back to the apartment: Nobody re-evaluates their vacation home every evening—and they don’t call you about it either. But with a public-markets portfolio, they do. With venture capital funds, the situation is effectively similar. There’s an extreme asymmetry between the book value and the actual daily market value, because the last valuation may well have been years ago—namely, the last funding round of a portfolio company. Publicly traded stocks, on the other hand, are priced in day by day. This asymmetry is interesting and triggers entirely different dynamics.

---

Jan Thomas: And this Griffin—had you heard of him before? He seems much more likable than I expected. When I read about that move, I thought, “Okay, he’s a real wolf.” But he comes across as likable, and I have no idea what the relationship between the two is like now. Do you think he sits there in the evenings and revels in the fact that he pulled the wool over Leopold’s eyes? Or is it almost a partnership-like coexistence? It could also turn into a sort of succession scenario—he’s around 60 now, I think—where at some point he says: I’m basically grooming the next generation.

Björn Rieckhoff: When it comes to financial investors, I can actually imagine it being relatively unemotional and level-headed. If you approach it with a clear head—of course, you can get nervous, and the phone is ringing off the hook, as you described earlier—but I do think that’s important, because otherwise you simply can’t handle these volumes. You’re emotionally detached from these figures and approach it completely rationally: Okay, I’m simply in a situation where the banks want their money back. Either I pay them, or we sell—which causes me problems—so I’d rather just pay. And the deal you end up with is, in the end, a fair deal—assuming it is fair—but it’s not the best possible deal he could have gotten for the portfolio. Of course not, because he’s in an emergency situation, and the others know that too.

Jan Thomas: And that must have happened within a matter of hours, right? The clock was definitely ticking the whole time.

Björn Rieckhoff: And then, of course, the question is: Are you mad at the person who gives you the money, or are you more grateful? I think it’s a bit of a love-hate relationship. You can break that down on a personal level, but if you look at it from a distance, you’d probably say: It was valuable to have this business partner, and when in doubt, there will likely be more business between the two of them in the future. As you say: The age difference is significant, and their levels of experience are quite different. All things considered, it’s a good learning experience for Leopold Aschenbrenner that will help him act more effectively in the future. But I would think: Situations like this also happen to financial investors who are 50 years old and have spent their entire lives on the trading floor. I wouldn’t take him out of the line of fire—but I also think it’s an overreaction to attribute everything solely to a lack of experience in the industry.

Jan Thomas: I just thought that Ken Griffin could actually be seen as a kind of prodigy at Aschenbrenner—someone who understands the market of the future much better than I do. Somehow, that symbiosis just seemed to fit in my mind. There’s actually footage of the whole thing—I don’t know if you’ve seen the movie *Margin Call*; it’s almost exactly this situation, with Kevin Spacey and Jeremy Irons, I think. It’s a really worth-watching movie that fits the setting here quite well. So there’s no need to remake it—that’s what I’m getting at.

Björn Rieckhoff: Nah, you’re right, but this dynamic is, of course, a recurring one. I think what’s important to understand—a few takeaways: It’s still highly leveraged capital, which means you can’t be fundamentally wrong. And that’s exactly my point: I remain convinced that Leopold Aschenbrenner wasn’t wrong in terms of the substance of the investments he made. But market dynamics hit him at the wrong time, and then it was triggered by the margin calls. That doesn’t mean he won’t have tremendous success with this thesis in the long run. And as a side note: He recently invested again right away—400 million in Source Foundry, a Stanford spin-off. Five days after the fire sale. So he remains convinced of this thesis—and, in my opinion, rightly so. This isn’t the trigger for “the AI bet is going wrong”—it’s a market dynamic created by this extreme leverage, which led to a situation where he had to sell.

Jan Thomas: That probably makes timing super important, too.

Björn Rieckhoff: Exactly. And that brings me back to the next analogy with venture capital. In Germany right now, we have a situation where fund investors are essentially saying: Please finally liquidate the portfolio—we want to see returns. There’s virtually no DPI—only book values. And the fund managers, who’ve had to defend this position in their investor meetings for quarters on end and year after year, are naturally giving more and more thought to the potential sale of their portfolio. For now, they’re still in a position to actively manage this—if anything. But at some point, you’ll find yourself in a situation where, as a fund manager, you simply have to liquidate. Then there’s no more grand posturing about “we’re going to look for the best offer”—it becomes a dictated price at which you have to sell because you’ve been forced into that situation by the timing. You should never find yourself in that situation. I believe this is the next stage of development for successful fund managers in Europe and also in Germany: that they start thinking about a divestment strategy very, very early on. And also liquidate portfolio holdings when they say, “We’ve reached our target return”—not in the hope that something might still happen before the fund’s investment horizon expires.

---

Jan Thomas: Yeah—there’ll probably be some exciting discussions there again. But maybe just briefly, to wrap up the Situational Awareness case: Do you think the fund is coming back? There was also that investment I couldn’t quite make sense of. I feel like everything is happening so fast—the fund sale, the wedding, the new investment. It feels like it all happened over a single weekend; I can’t quite keep up with this pace. And here’s something else I’m curious about: I took a look at Adobe’s stock price, for example. It’s just crazy when a stock’s price moves in a direction completely opposite to what you actually expected for your fund. Do you think we’ll see this slump in SaaS stocks over the next twelve months?

Björn Rieckhoff: I think we’re already seeing that to some extent when it comes to multiples—even in the private markets. And I believe that trend will continue. However, I also think the media is exaggerating it a bit. Because what SaaS businesses often have is good predictability—and it’s not as if these companies are sitting idle. There are various examples of companies that have responded very well to AI and have transformed themselves from a position of strength to avoid falling into the innovator’s dilemma and being overtaken from the right. I believe that many SaaS businesses, including publicly traded ones, will naturally take a hit, but they’ll be able to secure a future based on very strong recurring revenue—without first having to go through a bankruptcy-level distress scenario and then eventually find a new business model.

Jan Thomas: Well, I’m asking because companies like Salesforce and others have roughly halved in value over the past year and a half. But Aschenbrenner has now bet that it’ll drop even further. And that makes me wonder if that’s really true—or if at some point we’ll have to admit that these companies aren’t fundamentally all that bad, that they’ll eventually get the hang of AI, and might even generate more revenue as a result—per user, per seat.

Björn Rieckhoff: My take is that he’s taking a very aggressive approach here—just as he continues to be a strong advocate on the buying side for the AI trend to keep going. We’ve talked about this in recent episodes—when we’ve reached the limit and when we’re going too far. And he’s doing exactly the same thing on the other side, too. I think you have to say: It really is a case of playing it by ear. At a certain point, I think you do end up taking some profits from that investment thesis.

Jan Thomas: Exactly. Great, Björn—a really exciting topic. I think we’ll be hearing a lot more about this in the coming weeks and months. It’s now in the media spotlight—understandably so—and this is probably a story that’s just getting started. Maybe just to tie this back to the beginning: I saw these videos of him ranting at the Green Party convention ten or twelve years ago. I found that fascinating—he comes from a completely different background; you’d never have expected what he’d become. Totally exciting.

Björn Rieckhoff: I think he’s had an incredibly eventful life for his age. He’s accomplished in his mid-twenties what others manage in 60 years. Just think about everything that goes into building that kind of network. Honestly, I think he’s probably laughing at the European media reports that are now focusing on the alleged sale of his fund. He’ll see plenty of opportunities and will certainly work toward playing a very major role in the financial industry in the future.

Jan Thomas: Exactly. And I’d say we’ll continue to leave the German perspective out of this—we don’t need it for this.

Björn Rieckhoff: Yes.

Jan Thomas: That was fun. Thank you!

Björn Rieckhoff: Me too. Thanks.

Jan Thomas: See you next time. Ciao!

---

About Björn Rieckhoff

Björn Rieckhoff is an independent advisor and business angel with nearly ten years of experience in early-stage venture capital. He helped build Cavalry Ventures as its first employee and later became a partner of the fund. Today he supports founders more directly with fundraising — sharpening their story, stress-testing business models, and setting up lean financing processes. With over 80 transactions and board seats from seed to Series B, he brings this perspective as a sparring partner for entrepreneurs.

About Startup Insider

Startup Insider is the industry portal for the startup scene in the DACH region. It covers news from all regions and industries, along with an overview of key players and events in the German-speaking startup world.

Links

• Situational Awareness – das Essay von Leopold Aschenbrenner
• Margin Call (Film, 2011)
• Die BIS warnt: Was Eisenbahnmanie und Dotcom-Crash über den AI-Boom sagen (vorherige Episode)