Limitless: An AI Podcast - Situational Awareness: The Full Story of Leopold's Collapse

Episode Date: August 4, 2026

Leopold Aschenbrenner’s massive liquidation event followed heavy leverage and losses in its AI infrastructure bets. Today, we discuss the July timeline, the sale of the public equity book to Citadel, and the remaining Anthropic stake.------🔒 Check Out Our Sponsor: LEDGER AGENT STACK 🔒https://developers.ledger.com/?utm_source=Audio&utm_medium=Podcasts&utm_campaign=Limitless------🌌 LIMITLESS HQ ⬇️EMAIL US:            [e-mail]:    https://limitlessft.substack.com/FOLLOW ON X:   https://x.com/LimitlessFTSPOTIFY:             https://open.spotify.com/show/5oV29YUL8AzzwXkxEXlRMQAPPLE:                 https://podcasts.apple.com/us/podcast/limitless-podcast/id1813210890RSS FEED:           https://limitlessft.substack.com/------TIMESTAMPS0:00 Leopold’s Rise and Fall2:41 The Unraveling Begins4:14 Margin Calls and Liquidation7:03 Citadel Swoops In9:13 Hunting the Position10:19 Aftermath11:25 Anthropic Saves the Fund13:44 Ken Griffin's History16:30 Was Leopold Right?18:41 Bear Case21:21 Recursive Self-Improvement22:25 Ledger23:24 Leverage Lessons23:59 Closing------RESOURCESJosh: https://x.com/JoshKaleEjaaz: https://x.com/cryptopunk7213------Not financial or tax advice. See our investment disclosures here:https://www.bankless.com/disclosures⁠Josh works with Anthropic as a contractor. All views expressed are his own and do not represent Anthropic, its leadership, or its affiliates. Nothing in this episode is investment advice.

I can't believe we're filming this episode. The poster child of Silicon Valley investing. Leopold Ashton Brenner has kind of been dethroned. Like, the dude got wrecked. As to set some context, 20 days ago, there's this kid in mid-20s, Leopold Ashenbrenner. We filmed many episodes about him. He was running the best performing hedge fund on Earth. It was up 1,600%.

He ran it up from a few hundred million dollars to $45 billion. And then the market learned two things. One, that he was on leverage. And two, that he was getting closer to his liquidation prices. So what does the market do? hunted those positions and got him liquidated only for one man to come in and swoop up the entire position of his fund in one single transaction. And now, as I understand it, Ejas, every single one of his public positions is closed. That's it. That's it's gone. It is gone. Over night, it happens so

fast. This is unbelievable. Yeah. And I think it's important to kind of like set some context, who on earth is this guy and like how all of this unravel. We're going to get into all of that on this show. But on your point of leverage, Josh, like, it wasn't just like any amount of leverage. This guy was 4x levered on the entire font. Like, 4X is crazy. So the nominal value that he was levered at was a hundred billion dollars. Do you know how much money you need to borrow to be levered to a hundred billion dollars on like a $45 billion book? It is just absolutely insane. A lot more than it should have been. Way more. Way more than he should have been. Okay, so who on earth is Liverpool-Ashrenner? I'm

sure you've heard this name, but just a brief kind of recap. Leopold Ashrenner, was the wee age of 23 years old, so this was two years ago in 2024, when he left or rather got fired from Open AI on the Super Alignment AI research team, and he decided to write 165-page essay on what he thought the next decade of AI is going to look like. And turns out he's the only guy that was bang on with every single one of his predictions. And people love the essay so much that he raised a fund, a small amount of around $225 million. And over the course of two years, he rode that up 1,600 percent to the tune of $45 billion. Now, it is one of the most impressive runs of any investor. But the fact that he did it at the age that he had with no zero trading

experience, by the way, is just phenomenal. And the fund was based on two main pieces. Number one, that the physical AI infrastructure was going to be one of the best investment. investment opportunities out there, compute, GPUs, memory, all those kinds of things. He was very early on the trade. He called the trend very early. The second thesis is applications, software, he was going to be short. He didn't believe that companies like Microsoft or whatever their software was going to be worth anything in a world where AI models can just absorb all of those things. That was the sure bet. But things started to unwind around six weeks ago when his thesis that had held strong for two years started to waver, market started to recede,

there was the global war that was happening, there was a few kind of like oversupplies happening in terms of funding, and things started to go a little rye. Yeah, it was messy. I'd say it took about 20 days to go from like absolute legend fully liquidated. It was like, this happened very rapidly. And it was kind of marked in a way by that July 10th ringing up the bell of S.K. Hynix. This was like this huge IP. PO moment where everyone was very excited and it very much marked that top tick in terms of where the market was for Leopold in particular. So that was early July. You can think like July 10th is when it started. We're now sitting here August 3rd recording this. You'll be listening to this August

4th. In this month, a lot kind of happened. And it starts with the memory trade. We know Leopold is very risk on with the memory trade, so much so that he was using, like you mentioned, four times leverage. And just a brief explainer for those who don't understand leverage, at four times leverage, a roughly 17% move against your portfolio erases about two-thirds of the equity. So for every one dollar you put down to money, you borrow three more, you control four dollars of stock. If you're down like 25%, I think he wipes the entire book. Like you're done. It's just a 25% move. Exactly, 25%. But the problem with this is that the people who are loaning your money don't want to lose the book. So they're going to start to claw back that equity

prior to it reaching zero. And that's when you see this cascading liquidation of events. and this was caused initially by the market sell-off in memory, and then it just went a little bit further and faster from there. You'll notice that there was this entity that goes by the name of Citadel that did flag earlier than the week that, hey, we think interest rates might actually increase. And what does that mean for the market? It adds further selling pressure. So there was all of this pressure downward on his positions, and because he was leveraged, it creates a lot more pressure on those positions, and it required him to raise more money. So in early July, around this time, there were rumors that Leopold was looking to raise a little bit more money for the fund.

I'm not sure if he actually got there. But basically, July 24th comes. He writes a letter to his investors admitting the damage. Word gets out that he has been damaged. Those rivals press his known positions. The brokers that he loaned his money from, they want cash. And then by July 30th, the whole book sells in one single block and he's fully out of the market. And that's what happens. And it was really this like unbelievably devastating thing I would imagine for the fund. Because it just happened seemingly out of nowhere. And everyone went from like, oh my God, this guy's a genius to, oh my God, wait, you just lost all of his money. Well, perhaps not all of it, but a lot of the portfolio got wiped out.

Absolutely. Josh, I feel like this timeline could play in a movie like the social network or something like that. So I'm going to spend. It's crazier. It is nuts. Like I saw this entire thing unravel on my timeline, right, in real time. And so I'm going to share some of the tweets that kind of, I'm going to take you guys. through this journey and run you guys through this entire timeline. So the original tweet that went viral

was news broken from the Financial Times. And the title is a little demeaning. It's Leopold Ashenbrenner's Situation Awareness seeks to raise capital after the AI route. Now, that was the, I think it was the start of July that you just referenced Josh, where it was like, okay, things are getting a little weird. And I watched the Martin Schrelli interview on TBPN, and he basically said he got approached by someone, like a random intermediary saying, hey, do you want to buy $100 million of anthropic shares at a really steep discount? And he just kind of sat back in his chair and he was like, is this Leopold? Because there's no other fund out there that would, you know, do this in their right mind.

And then from Martin Scralli himself, he goes, either hearing rumors that Salp, which is situational awareness, LP fund, is down more than 50% month to date after it was being up around 200%. Now, the reality is even worse. It was down more. It was down 67% according to Leopold's letter at that time, but it wasn't publicly released back then. And the sad news is month of a month, or rather year to date, he was up 440%. So like, you know, all of this came kind of like after the fact. So we're kind of like seeing a lot of these like news bits break. Like people are like, okay, do I think that this like fund maybe potentially broke off? And then the news came that you

just referenced Josh, which is Leopold sold his entire public equity book that was levered to this man called Ken Griffin. Ken Griffin is CEO of Citadel. Now, Citadel, they've been in this game for a while. They're our market makers. They're one of the biggest market makers in the world. And they are known for managing risk expertly, such that they give compounded, very heightened gains every single year. And Citadel was the one that situational winters or Leopold approached when they were in this time of need when their leverage was working against them and said, Ken, I need to save this fund. I need you to help me out of this very sticky position. So Ken's team worked with him for six hours into the early hours of the morning before market open on Monday last week and said,

okay, you can sell me this book at 40 to 80 cents on the dollar. And Leopold agreed, Monday market open. Remember that relief? Did you see those stocks? Did you see those stocks? There was a lot of green. It was nice to see a lot of green. They were up like 20% apiece. Insane. So the memory stocks that had been beaten down over the last month were up like 20 to 25%. Guess how much Ken made in a couple of hours just by buying Leo's book?

Many billions of dollars. Yes. He made four to six billion dollars if rumors that were stated are true. So a very interesting thing. But Ken, of course, he's a pro at this game. And he was already working against Leo before this happened. If you remember that Ken was pushing. for interest rates to get like even worse, even though he probably knew what the outcome was,

just to stoke fear in the market and dump the markets even more. So people became aware that Leopold's fund was in trouble because of the recent decline. And like you said earlier, they all piled in. They basically all wanted to short the position, push the stock down even more so that they can liquidate the guy and get in at a better price. And that's exactly what King Griffin did. This is like an episode of billions, if anyone's watched the show. of like Bobby Axelrodd is Ken Griffin. And you saw this predatory selling happening in real time. And that's when the market knows that there's this giant fund wounded. And it also knows the positions. If you realize, I mean, every single quarter, we talk about the 13F filing. That shows the positions of the fund. It knows whether to sell or short those names and then force it out and then buy it back cheaper. So people were able to actually target specifically the positions that Leopold held in order to push those liquidation rates down higher, down faster and force him to sell all this thing. And then the lender's call. I know his prime brokers,

who are the people that lend him all the money, these are companies that you know like Bank of America, Goldman Sachs, J.P. Morgan. They call him up and they say, hey, those stocks that we loaned you, they're down a lot. We're going to need that money back right now. So he has to sell it. And on top of that, the rate hike thing is so funny because you have to imagine Citadel is playing this game. They were targeting the positions that Leopold had to sell them to push the price down. And then adding the rumor on that these interest rates are going to be increasing. That crushes the market even more. It's like, oh, man, that's so brutal, only to buy it back at the absolute bottom and run it up to, from what, you've made it up like five, six, seven billion dollars in a single day. So I'd love to hear

the behind the scenes of how all this happened. I think that's probably really funny. But if we look at this book before and afternoon, we can kind of digest the carnage and see where he stands now. Like, what does the situational awareness fund look like? Well, it turns out that almost all of the public positions were gone. There still seems to be a few small ones, but the main survivor is the anthropic shares. Funny enough, it sounds like they still have about $5 billion in anthropic shares. So now the situational awareness fund has become a holding company for Anthropic. Is that right? Yeah, pretty much. It's a venture fund? It's a venture fund right now. Martin Schrelli had this crazy take on CBPN where he basically goes, yeah, when you see a hedge fund starting to invest in

VC deals. It is the death knell. And literally a month later, like, Leopold's fund like blows up. But yeah, to kind of give you the state of the situation right now for Leopold, his public book is pretty much entirely erased. He has a few positions open. And they're completely unleavened and they're going to be unleathered for the foreseeable future. So it's spot only, long only. And then, yeah, the Anthropic shares is actually what ended up saving the fund. So this is a unique twist. And I'm going to show you, um, actually, From the horse's mouth himself, Leopold Aschenbrenner's letter that he shared with his investors, I'll give you the summary of what he says in the letter, but basically he said,

we had an extenuatingly bad month, and we let you guys down and it's completely my fault. So in a month, the fund drew down 67%. However, year to date, the fund is still up 80%. So I just want to repeat that for a second. The fund drew down massively, 67%. but was still reported an 80% gain year to date. The only sole reason why they were able to report that is because of this little baby right here. Oh, boy.

Post for child. Correct. So he sold a huge amount of anthropic shares, and he got in super early. I think it was in their series 8 round, so when the company was valued at $60 billion, and they are now, like, I mean, if you believe some of the secondary markets worth like $1.2 trillion or something like that. So he made a huge bab on that and he used that to basically quell the huge public market loss in that. And in this letter, he basically admits to his fault, but he says that these are extremely expensive scars that he needs to learn from. And the most important thing is he gets to live another day.

Now, if you want to understand, you know, which investors are kind of okay with this and which investors might be bad, the early investors are probably okay with this definitely hurt. They're licking their wounds, but they're like, okay, I saw Leopold run this up. His thesis is still intact. So let's see what he could do. But if you were part of that capital call, which you mentioned earlier, Josh, which was like just before all of this was going bad, when he was reporting like, hey, we're up 440 percent and we're taking on more capital, right? Because the markets were starting to draw down.

If you put money at that point, so that was like a month and a week ago, you're gone. Like your entire LP share is erased. So that's what's not being spoken about publicly. And I just wanted to kind of lay that out. It has been a brutal turn of events. And the worst part is he had his wedding. this past weekend. So he was going into his wedding whilst all of this was happening. That sounds like an absolute name. I really feel for the guy. This has got to be such a

tough learning experience. But I mean, in a way, like you made it to the big leagues. Like he ran up this fund from a couple hundred million dollars to 45 billion and now you're going to have to fight with the big dogs. And Ken Griffin has done this before. He did this. What was the oil company? Enron, I think it was that went out of business. It was unbelievable. So Ken Griffin is actually worth highlighting here in the story because he is, I mean, a remarkable bailout investor in a way, Ken Griffin is the type of guy that will never get a phone call, but perhaps like once every five to ten years he'll get the phone call and it will make him tens of billions of dollars every single time. He's the guy that you could call to bail them out. If anything goes wrong,

if you were on the edge of bankruptcy, you call up Ken, he'll send his guys over and he'll take care of it. I know with the Enron story, at least. I remember this story because it was so amazing, how he took a bunch of his top investment guys, put them on a plane and flew them over to go actually be at the office and work overnight to process all the books to see where the value was in the business. And what he discovered is that a lot of the value was in a few key people that were kind of managing the infrastructure that truly understood the business. So he extracted those people. He started his own thing with them. And then the rest of the company was kind of sold off to someone else. And it was having a really tough time. It didn't do well. But Ken and the team, I mean, what does it say?

The team built Citadel's commodity trading and made $30 plus billion so far where the UBS who bought Enron. They had to shut down Enron, which was the business that had bought. So Ken Griffin has done this before. He is the guy that you call. Leopold called him. It seems like he's always kind of at the center of some sort of, you know, interesting cultural moment. I remember the last one for the people who were involved in crypto. It was the, um, oh, what was this? It was like the Nash. GameStop. GameStop, not game. It was game stop partially. And then there's the other, the Constitution. Amara. That's, oh, the Constitution. Yeah. Yeah. See, he's part of so many things. All these cultural moments, it's like, okay, the GameStop moment, he was in. The Constitution, he was in. Like, when people were buying the Constitution, it's just like, Ken Griffin is always there lurking in the shadows and capitalizing on this. So I think that comes in, I think when it's like, I don't know that I think about it. It's like, yeah, it's like the GameStop thing when Capital was getting squeezed, he was like, all right, I'll bail you out for $10 billion. But like, I think he got recurring percentage of their revenue from that fund for the immediate future and like for the long foreseeable. future as well. So he just like had passive income coming from this like billion dollar plus

fund. And then the other one I think I remember was the Amaranth thing. This was before my time. I think I was like a little baby, but I read up about this. And it's this guy almost Leopoldesque. And he was trading gas futures, Josh. And he made like a similar return back then. And then he flew too close to the sun, blew up and sold his entire book to Ken Griffin. So actually, I think Ken has made the most money from these disaster situations. Just a shock. Yeah, he's he's top dog. I mean, people are going to quickly learn when you get to that size. Like, you're going to have to deal with the big boys. You're getting calls from Javan Morgan.

You're getting bought out by Citadel. I mean, this is the reason why these behemoths that exist today. So was Leopold wrong or did he just get wrecked? Look at this shot, Josh. You tell me. Look at this shot and you tell me. That's so brutal. Okay, so upon liquidating his entire position, all of his holdings were up an unbelievable amount.

This is one day, right? This is one day. So Nebius, one of his largest holdings, up 27%. Iran, up 26.5%. Bloom Energy, which we filmed an entire episode on, up 25% in a single day. Look at this. Look at this. Look at the salt to wound that clothes adding to this by midday.

It came up 16%. So his entire portfolio absolutely ripped, which signals to me. And I mean, based on everything that we've been saying, too, I mean, the day before this happened, we filmed an episode saying, the market's wrong because we were watching the sell off. And we were trying to make sense of why the market was selling these things off so dramatically. And I think now we kind of have an answer. People were hunting Leopold's positions. As soon as he got liquidated, as soon as those positions closed, the entire market ripped. And I think it's a testament

to Leopold in the fact that he was right. It's just he was right with leverage. And unfortunately, with leverage, you are never truly secure. You are never truly safe. Even if you're right, if you're right in the wrong way, it's just as equal to being wrong. And that's kind of where he fell here. But I think directionally, he is right. And now we have to reevaluate this question. like, hey, is this a good time to actually deploy money into these companies? Because, look, they're doing well. We just saw all of their earnings reports from companies like Google, whose cloud margin revenue is going through the roof. Their CAPX is going through the roof. They're so much that they're cash flow negative for the first time in company history.

So all these large cap companies are spending huge amounts of money. We know where that's going. It's going to land power shell. It's going to the chips. It's going to the memory. It's going to all the infrastructure required to build these tokens. And who are the people that are responsible for this? Well, we're looking at the list. on screen right now. So it seems like Leopold is right. He's probably going to try to run this back as best he can. I'm sure people shouldn't lose too much trust in him. I mean, that's like a touchy thing to say, but he wasn't wrong. He was only wrong in terms of how he went about it. Hopefully this is the learning experience. And as we move forward, the thesis still stands and it

can kind of continue along this journey of being that like poster child for the AI investment trade. Yeah. I want to give the other side of the coin here, which is like, What if Leopold is wrong? And there is a convincing enough argument that he might be. And this is not something I prescribe to, but I want to give that for the bears that are watching this show, right? Okay. So if you remember earlier in this episode, I said his thesis were two parts. One, that AI physical infrastructure was going to keep going up because the demand is way higher than anyone can conceivably think of.

And number two, that he was going to be short software applications because AI model companies or labs like Anthropic and Open Air are just going to, absorb them, right? They're just going to get the model to train and do the thing that Microsoft's application can do and then just replace Microsoft, right? But like you just said, Microsoft just had their earnings report and it is the best that they've had. It's a record earnings investment for goodness knows how long at this point. And that's been a continuing trend across most software applications that Leopold was actually short in his most recent 13F findings, cybersecurity stocks and a number of other ones. They are actually all up over the last couple of months. you know what hasn't been up over the last month, up until maybe like market open of like last week.

It's these memory stocks, man. It's the AI physical infrastructure. Invidia actually is up 0.33% over the last month and a half. So like there's a lot of things that are going on here that could potentially hint that Leopold's thesis is wrong. That being said, I don't think that that is correct personally, purely because of all the demand that the like. So if Microsoft, Amazon, hyperscalers are seeing on the cloud service side of things, that's going to drive more memory demand. It's going to drive more GPU sales from Nvidia, from AMD. It's going to drive more CPU sales from Intel. The infrastructure play is very much still there. Now, the question is,

is it already priced in? And that honestly is something that I can't answer, because I don't know what people have invested in or like how much of their money they've invested in. If you look at our friends in Korea who are housing two of the memory giants, they're all leverage up. They're all borrowing from their banks. So I don't think we've quite seen that extent here in the West, but I do think we are in a position where like, it could go up, it could go down, I don't know. But over the long term, I do believe infrastructure is still very much in demand

and Leopold will ultimately end up being right, which is a very, very expensive lesson to learn on your wedding weekend. And I hope that the guy makes it out because he's still, he's 25 years old. And I've seen a lot of people like hating on him. And like, listen, I understand I get it. He's lost a lot of money. It was very irresponsible. But to pull something off, to have the returns that he had,

he's still 80% up on the air. If you want to kind of take that number at face value, he's beaten a lot of the traditional headfront still, right, after this entire drawdown. And I hope he learns from it. And I think that he probably will and his thesis will play out. The next thing is RSI recursive self-improvement. Josh, maybe we need to do an episode on that, I think. Yeah, we're going to talk about that. And many other things. There's a lot of moving pieces now that are happening. And when we look at the market, it's like, we don't really know where things are going. So the best you could do is guess. And we've seen these guesses with like many, many multiples on the revenue of these companies.

It's starting to compress a little bit. We're starting to see that because of the. uncertainty. A lot of these companies are at capacity in terms of the bandwidth that they can create. So the only surprises can come really from the downside as opposed to the upside. And there's a lot of these like market forces that are at play that are pushing against this thesis, at least in terms of the memory companies, the AI companies. Like when you think about China, they're getting much closer. They're starting to turn these tokens into commodities. If they do, they're fighting an energy war with the US instead of an intelligence war with the US. That makes things kind of slightly in their favor. But I do.

agree on like the infrastructure trade at least we need so much more power we need so much more memory we need so much more tokens and assuming that continues to hold true um you got to assume that leopold would be there so like nothing but respect for the guy sucks uh that's a bummer sorry to hear that you know what could have helped leopold is if he was perhaps using i mean ledger have you heard of this because if you are building with a i agents you're probably worried about security and an agent with unchecked access is a problem um you can think of a portfolio with leverage is a problem types of problems, except ledger is protection to help you solve this problem. Ledger lets the agents propose. It lets humans approve and then ledger signers enforce. There's a three-step process that works

with cloud code, codex, cursor, it's open source, it's available today. It works with all the things that we work with. There's basically this thing called the ledger agent stack, and it fixes this using open source tools that allow you to engage with agents and then tell them exactly how you want things done, approving them along the way. So thank you so much to Ledger for sponsoring this episode. Hopefully, can find his own version of Ledger. And yeah, you can find the link of the description down below at developer.com. Can we end this with a meme? You got one?

Let's see. Por favor? Yeah, I got one. We named the fun situational awareness, but lack the situational awareness to sell when we were up 2,200% in two years. That's pretty good. That's pretty good. Some of these memes are so good, dude.

It brings me pain because it's like, yeah, dude, obviously. But also, I mean, I would stay risk on too. Just with less leverage. Don't use leverage. Everyone knows. The lesson here is stay away from leverage, guys. Or if you're going to use leverage, certainly don't do 4X on a $45 billion book.

What are you doing? You don't need $100 billion. Just you're already one of the best performing headphones in the world. Like, just chill out. But we will keep track of everything that is unraveling. When Josh and I filmed the last episode covering the Leopold story, which was literally a few days ago, the news was breaking as is.

And as we wrapped up recording all of this stuff came out. So we felt the whole episode. We were like, we need to do another episode. So this is that episode. So if you enjoyed it and you are listening to this on YouTube or Spotify or Apple Music, wherever you are, please give us a rating. Leave us a comment. It helps us out massively. We've been hearing from a lot of you. I got accosted in the street, Josh, from another fan that walked up. Dude, that's crazy. Well, he goes, hey, you're the, you're that podcast guy. Like, you were speaking about this episode. We're making moves. But I do say, like, if you see us in the street, like, we would love to say hi to you. Like, we don't know. We haven't met too many of our listeners that. Um, that are. I guess extended family and friends. So please come out to us. Say hello. Leave us a comment if you would like. It helps us out massively. Is there anything else? We're covering both coasts now. I came back from SF. EGEN is now in S. We're just like we're going by coastal or we're getting the show done

either way. If you did enjoy this, don't forget. The one of the most important things you could do is give us a new listener. Share it with a friend who might also enjoy this. That really goes a long way. And then yeah, if you ever do see us, say hi. I mean, that'd be pretty cool. It's fun to see the numbers on the screen translate people in the real world. So that's always a really good time. I hope you get recognized more. You should just like, I don't really, I'm not super familiar with SF, but like you should just go to like the hot spots, EJazz and just kind of sit there and try to ORAFarm. Like, let me know if anyone comes up to you and says hi and just like report back how many fans we actually have. Hey, maybe I'll see Leopold at a cafe at this point, dude.

I think the dude's down bad. Like maybe, maybe he'll be down to have a conversation. Well, wish him well for me if you do because I really hope he pulls it together. I hope that this fund manages to claw its way back, as I'm sure it will. And yeah, that's Leopold's story. So thank you all so much for watching as always, and we will see you tomorrow on the next one.