4x Leverage, Long/Short Carnage, and a Hundred-Billion-Dollar Hunt: The Full Story Behind the 25-Year-Old "AI Stock Whisperer's" $45 Billion Fund Blowup
- Core Thesis: The collapse of Situational Awareness Fund (SALP) was rooted in roughly 4x leverage combined with highly concentrated holdings (including private equity in Anthropic), creating a fragile structure where a 25% drawdown meant total wipeout. Market traders accelerated the blowup through a "hunting" mechanism; macro narratives were not the primary cause—marginal trader behavior and leverage levels were the decisive variables.
- Key Elements:
- SALP employed approximately 4x leverage, with portfolio market value dropping from $120 billion to $90 billion (-25%), triggering a net asset value wipeout. The $35 billion in principal likely shrank to $500 million, with prime brokers Goldman Sachs and Bank of America taking forced control and disposing of assets.
- Market hunting mechanism: Once traders caught wind, they sold overlapping positions and shorted the fund's holdings, with Wall Street's Darwinian dynamics accelerating the collapse. Citadel acquired the assets at a discount as a package, potentially booking $3-4 billion in paper gains.
- Private equity is a "kiss of death" for highly leveraged hedge funds: The Anthropic equity stake could not be liquidated at the push of a button, and Leopold's deep personal ties to the company's leadership (through his fiancée) created overlapping concentration and liquidity risks.
- Missing track record: Leopold worked at FTX until its collapse, returning two years later with higher leverage and even more concentrated positions. His late 13F filings reflected operational immaturity, and institutional investors who rejected him for "lack of experience" proved prescient.
- The Kelly criterion lesson: Even with a 60/40 win probability edge, if position sizing exceeds the optimal level by 2-10x, simulation results inevitably trend to zero over the long run. Historical cases (LTCM, Amaranth) all validate the lethal combination of leverage and illiquidity.
Compiled & Edited by: Odaily TechFlow
Hosts: John Coogan & Jordi Hays
Guest: Martin Shkreli (Shkreli previously worked at Cramer's hedge fund, later founded multiple hedge funds including Elea Capital and MSMB Capital, as well as biopharmaceutical companies such as Retrophin and Turing Pharmaceuticals. He personally experienced the 2000 dot-com bubble and multiple institutional blow-up cycles)
Core Topics: Post-mortem of the Situational Awareness Fund (SALP) liquidity crisis, hedge fund leverage mechanics, prime broker forced liquidation process, and market hunting psychology
Source: TBPN Podcast --《Martin Shkreli Breaks Down the Collapse of Situational Awareness》
Airdate: 2026-07-30
Conflict of Interest Disclosure: Martin Shkreli is currently an active individual investor. He mentioned on the podcast that he holds some AI-related stocks (including Kosha in Japan). He also runs the pharma-tech company DrugDash and operates a paid subscription business. The views in this episode are based on his Wall Street trading experience, but his personal holdings may align with the direction of his interests when discussing specific targets.
Incidentally, Martin Shkreli is the "financial bad boy" who single-handedly shook the US pharmaceutical industry and Wall Street, was charged with manipulating pharma stock prices, and was sentenced to 7 years in prison. Because he himself is the "godfather" of playing with leverage, short selling, and gambling against regulators, his breakdown of this $45 billion AI hedge fund blow-up offers a cynicism and understanding of Wall Street's law of the jungle that far surpasses typical financial commentators.
Key Takeaways
1. The Situational Awareness fund used approximately 4x leverage, meaning a 25% drawdown would be enough to wipe out the fund's net value. When the portfolio's market value fell from $120 billion to around $90 billion, the $3.5 billion in equity may have shrunk from $3.5 billion to $500 million, triggering a forced takeover by the prime broker.
2. Market participants were "shooting against the fund" as early as Monday and Tuesday this week—selling positions that overlapped with the target fund and shorting its holdings to actively accelerate its collapse. This is classic Darwinian Wall Street behavior.
3. The three major bidders for the assets were Jane Street, Millennium, and Citadel, with Citadel ultimately winning. Shkreli believes Citadel could realize $3-4 billion in immediate paper gains after taking over, provided they can smoothly digest these positions.
4. Shkreli argues that macro narratives (war, oil prices, open-source anxiety, peak hyperscaler capex) are all noise. What truly determines prices is the buying and selling appetite of the marginal 5% of traders, and the leverage they're carrying.
5. Leopold previously worked at FTX until the day before it collapsed. He should have learned risk management lessons, yet re-entered the market less than two years later with higher leverage and a more concentrated approach. Additionally, he broke the hedge fund convention of not investing in private equity by heavily allocating to highly illiquid assets like Anthropic.
6. A major New York fund-of-funds reportedly turned Leopold down during fundraising, citing "no experience, can't invest." Leopold later delivered 20x returns, leaving them embarrassed—but after this blow-up, they have been "somewhat vindicated."
7. Shkreli uses the Kelly Criterion to illustrate: even if you have a 60/40 edge, as long as your position size is 2-10x the optimal size, simulations always end in going to zero.
Highlights
"I was talking to friends about Long-Term Capital Management, Amaranth, and other famous liquidity-driven blow-ups. The Situational Awareness situation definitely ranks up there."
"Once the market knows a fund has to liquidate, the most advantageous move for everyone else is to sell overlapping positions and start shorting everything it holds. It's brutal and Darwinian, but very common on Wall Street."
"None of that is the core. What really matters is the buying and selling disposition of buyers and sellers. Smart money gets in early and keeps buying as prices rise. Then less smart people see the gains and want in. The weakest hands tend to buy at the top and are the first to panic-sell. Every bubble is roughly the same: euphoria, peak, then everyone panics at once. Fundamentals barely matter at times like this."
"Ken wants to be the guy everyone goes to when they're in trouble. Buffett is old and doesn't want this kind of work. But Citadel did exactly this during the Amaranth blow-up."
"When a hedge fund puts on a venture capital hat and does private equity, it usually doesn't end well. Looking back over 50 years of hedge fund history, very few people have been able to do both well simultaneously."
"Leopold didn't do anything wrong—the leverage level had already determined the outcome. At the slightest disturbance, he was going to blow up. There was no other ending. It's unfortunate."
"If you're going to hold these stocks, you need to make sure you can hold them without flinching when they drop to 2x or even 1x P/E. The only people in the world capable of holding $100 billion without flinching are probably at the level of Citadel."
Full Transcript
1. Opening: The Wildest Month in Wall Street History
Shkreli opened the podcast by saying the past 24 hours were among the craziest experiences of his personal investing career. He mentioned that he and friends in the industry had heard rumors about the Situational Awareness fund running into trouble since mid-last week, with the picture becoming clearer by Thursday night and Friday morning. He placed this event alongside famous liquidity-driven blow-ups like Long-Term Capital Management (LTCM) and Amaranth, calling it "definitely up there."
What impressed him was that the fund did a fairly good job with secrecy. But the market has a keen nose, and some large counterparties may have started positioning as early as Monday or Tuesday this week. He quoted his former boss Cramer, calling it "shooting against a fund"—when a fund is forced to liquidate, the optimal strategy for others is to sell overlapping positions while shorting everything it holds. It's not a moral issue; it's pure game theory.
2. Root Cause: Not War, Not Oil—Marginal Traders and Leverage
The hosts threw out a series of macro narratives to test him: US-Iran war? Oil prices? Open-source AI anxiety? Peak hyperscaler capex? Shkreli dismissed them one by one.
"None of that is the core. What really matters is the buying and selling disposition of buyers and sellers." He described the classic bubble psychology pathway: smart money enters early and keeps buying as prices rise; latecomers see 400% returns and follow in out of FOMO; the weakest hands buy at the top and are the first to panic-sell. Shkreli joked, "People like me start buying near the top. I think memory is great, and bottleneck trades are great."
He emphasized that fundamentals barely matter at times like this because prices are determined only by the marginal 5% of traders. And the problem is precisely that these 5% are operating at 3-4x leverage. According to market rumors, SALP used roughly 4x leverage—"a 25% drawdown takes you out."
3. The Math of 4x Leverage: From $45 Billion to $500 Million
Shkreli used a simplified set of numbers to help listeners understand the brutality of leverage. Assume the fund had $35 billion in equity, plus roughly $10 billion in Anthropic private equity (as understood at the time), bringing book equity to about $45 billion. Running at 4x leverage means total position market value was approximately $120 billion.
When position market value fell 25%, from $120 billion to roughly $90 billion, book equity shrank from $35 billion to around $5 billion, or even lower. Once equity approached or fell below zero, the prime brokers (Goldman Sachs, Bank of America, etc.) stepped in. They're not there to save you—they're there to take over assets and sell them as quickly as possible, because "their boards would rather be certain of losing $1 billion than risk losing $5 billion."
There were rumors that Leopold urgently contacted about 10 institutions over the weekend trying to sell Anthropic equity to raise liquidity, with offers based on an Anthropic valuation of roughly $1.1 trillion. But ultimately, disposal of the public book fell to the prime broker, and Citadel stepped in as the buyer, taking over the chess game at an overall discount.
4. The Hunting Mechanism: When the Market Smells Blood
Shkreli detailed the practical difficulties of unwinding a large position. You can't handle a $100 billion position like clicking "sell" on Robinhood.
The normal process is to call Goldman Sachs, who acts as an intermediary to find buyers. But the intermediary is obligated to "advertise" the order to the market—publishing their market maker identifier (like GSCO) and the assets for sale. Once the news spreads, all of Wall Street knows there's "a big seller."
Several things happen then: smaller funds might quietly short the asset, trying to get in front of the large seller; institutions that genuinely want to buy also hesitate—"if his inventory is really that big, I need to be cautious getting involved." The holder list is only so long. You call Fidelity, you call index funds, and they all say they haven't sold—"then it has to be him."
Even more brutal: when the market confirms someone must sell $100 billion, "trillions of dollars will stand in his way, just to watch him cry." And this isn't just Leopold's $100 billion—multiply it by 5 to 10 times to get the total capital in the market making the same trade. Shkreli believes that while the most violent liquidation phase may have passed, more funds will be revealed with 30-40% losses over the coming weeks.
5. Citadel's Entry: Ken Griffin Wants to Be "That Guy"
During the asset bidding process, Jane Street, Millennium, and Citadel were brought into a closed circle. Shkreli heard that Millennium did bid, but Citadel's terms were better.
He assessed Ken Griffin's motivation: "Ken wants to be the guy everyone goes to when they're in trouble." Buffett is old and doesn't want this kind of mess, but Citadel has played similar roles in past crises, including the Amaranth natural gas blow-up and the Enron collapse. This is an extremely expensive brand investment—"might be used once every ten years, but when it is, it earns $5-10 billion for free."
Citadel actually posted a small positive return this month, which Shkreli believes is likely because they were already hedged. More importantly, as a major prime brokerage client, Citadel holds vast amounts of global trading volume data, giving them a natural advantage in information and execution speed.
6. Your Prime Broker Is Not Your Teammate
Shkreli explained the prime broker business model: they make money on financing spreads. If you borrow at 4x leverage, the prime broker might capture 400-800 basis points of "free income." So they love leverage.
But the prime broker's internal risk department is monitoring a different set of metrics: excessive concentration is not allowed, and overly large short positions are also problematic (the GameStop lesson). What gives them the biggest headache is private equity. In Shkreli's view, a hedge fund doing venture capital is a "kiss of death." East Coast hedge funds doing private equity typically can't compete with West Coast dedicated VCs.
Leopold's situation was especially tricky: Anthropic was a private equity holding, and he was as close to the company as humanly possible (his fiancée is the Chief of Staff to Anthropic CEO Dario Amodei). Although Anthropic's demand had surged 100x over the past six months, when you need cash, "you can't press the sell button."
Rumors say someone tapped Leopold on the shoulder Monday or Tuesday saying, "Your margin looks a bit thin—can you top up a few billion?" But things moved too fast; there simply wasn't time.
7. Can Leopold Make a Comeback?
The host asked whether Leopold could rebuild his career. Shkreli believes he absolutely can.
He cited Peter Thiel as an example: Thiel's macro hedge fund Clarium Capital underperformed in its later years, but he then pivoted to Founders Fund, becoming one of the most successful VCs in history, and later restarted Thiel Macro. Shkreli said Leopold could take a few years to rebuild and learn from his mistakes—"nobody denies he's a genius."
But the process will inevitably involve humiliation. Two months ago he ran the world's largest hedge fund; two months later he's forced to liquidate—"this is an extremely crushing moment." Additionally, clawback clauses in the hedge fund industry could complicate things. Many institutions now require fund managers to return previously earned 2% management fees plus 20% performance fees in the event of severe drawdowns.
Shkreli also noted that Leopold's late 13F filing had sparked market speculation—"everyone thought he'd negotiated some NDA, but it sounds like he just didn't get around to it." This reflects the operational and communication immaturity of a young fund.
8. Lessons for Everyone: The Kelly Criterion and Position Sizing
At the end of the podcast, Shkreli shared his own position size simulator based on the Kelly Criterion.
The Kelly Criterion proves that if your win-rate edge is 55%, the optimal position size is 10% of capital. But in reality, almost every trader is betting at 2-10x the optimal size. He demonstrated with his simulator: even with a 60/40 edge, as long as you over-bet, the result is always going to zero.
He recalled that after leaving Tiger Cub, he had the opportunity to observe a low-profile fund manager who had worked at SAC Capital (now Point72) for many years. This person managed $300-400 million, almost entirely his own money, with 80-90% of capital sitting in cash year-round. He only made small trades, never had a losing quarter in over 20 years, and delivered 20-30% annualized returns.
"And the first thing I did when I got capital was to go 8x leverage. The dumbest thing in the world."
Shkreli said this ultimately comes down to psychology. Hedge funds are the sexiest, most painful, most terrifying business in the world. You think you're the master of the universe, but in reality you're waking up at 3 AM to check Korean stock prices, then again at 6 AM to see what happened in the world—"you're basically doing nothing but playing a high-stakes game of crazy poker."
