The "greatest legend" of this bull market comes to an end! In just a few weeks, the "AI stock god" is defeated, Citatdel acquires all positions

A 24-year-old, with a long essay predicting that AI would change the world, managed to raise $45 billion on Wall Street and achieved a staggering 439% return in the first half of this year. However, in just one month of the recent AI sector’s crash, he lost all the mythical gains accumulated over two years in one go.

On July 30, a message circulated on Wall Street:

An eight-employee hedge fund in the past 36 hours liquidated its entire public market position—about $16 billion in stocks—selling the whole package in one block trade to a counterparty.

The buyer was Citadel, one of the world’s largest hedge funds, founded by Ken Griffin, with $71 billion in assets under management.

This forced-to-liquidate fund is called Situational Awareness (SA for short), founded by Leopold Aschenbrenner, aged 25.

Of note, Aschenbrenner’s fiancée, Avital Balwit, is chief of staff to Anthropic CEO Dario Amodei. This AI crash and his forced liquidation coincided exactly with the weekend before their wedding.

And just a few weeks prior, he was still the most sought-after name on Wall Street and in Silicon Valley.

The “AI Prophet” and “AI Stock God” Worshipped by Silicon Valley

Leopold Aschenbrenner, German, 24 years old.

Admitted to Columbia University at 15, graduated top of class at 19. After graduation, joined OpenAI’s “Superalignment” team, focused on how to align superintelligent AI with human values. In 2024, OpenAI fired him for “inappropriately leaking internal information.” He denied this, claiming he had only shared a basically non-sensitive planning document with external researchers for feedback; the real trigger was an internal memo he wrote criticizing OpenAI’s safety measures.

After being fired, he did something nobody expected—he wrote a 165-page essay titled “Situational Awareness.”

The core assertion of the article: AI is developing far faster than people expect. By 2027, AI will be able to independently conduct AI research; humanity will face true superintelligence. All this will require explosive expansion in chips, memory, data centers, and power infrastructure.

This piece caused a sensation in Silicon Valley. Some called it “the most important article of the decade.” Podcast host Tim Ferriss dubbed him “the Nostradamus of AI.”

In 2024, he launched the eponymous SA fund with around $225 million in seed capital. Investors included Stripe co-founders Patrick and John Collison, former GitHub CEO Nat Friedman, and institutions such as Jane Street.

With no prior industry experience, he immediately managed billions of dollars. The fund quickly ballooned, reaching nearly $45 billion by early July 2026.

The entire fund had only eight employees, including four investment professionals.

Heavy Bets on AI Infrastructure, with 4x Leverage

SA’s investment logic was direct: since AI requires massive computing power and infrastructure, heavily invest in companies providing computing power and infrastructure.

Long positions: SK Hynix (Korean semiconductors/memory), Nebius (AI cloud platform), SanDisk (storage chips), Micron (memory chips), CoreWeave (AI cloud computing), Bloom Energy (energy).

Short positions: Shorting traditional software companies like Adobe—the rationale being that AI will disrupt them.

The fund further magnified returns by borrowing from banks (using leverage), up to about 4x leverage.

This structure has a feature: banks hold physical shares, while SA uses a “total return swap” (TRS) to synthetically obtain economic exposure and leverage. The benefit is amplified returns, but the downside is if the stock price falls, banks will issue margin calls.

From 2025 to the first half of 2026, this methodology worked wonders.

By the end of June 2026, SA’s net rate of return hit 439%. The fund grew from under $1 billion at inception to a peak of about $45 billion by early July.

Some early investors achieved up to 200% returns this year.

On July 24, Aschenbrenner was still writing to investors: “Sometimes we specifically point out that now is a particularly good time to add—if you’ve been waiting.” In the postscript, he invited investors to add new capital on August 1.

Hit on Both Sides Within Three Weeks

However, good times didn’t last. In July, AI stocks began to fall.

Markets started questioning: Can the high valuations for AI infrastructure persist? Huge amounts of capital exited AI hardware stocks.

SA’s long positions were hit hard:

• Nebius plunged about 48% from its June peak, evaporating $35 billion in market cap
• SanDisk fell over 56% in just over a month
• CoreWeave, Micron, and SK Hynix all dropped over 35% this month

At the same time, the software stocks SA was shorting (like Adobe) moved against them and rose, causing losses on their shorts as well.

With losses on both long and short ends, and 4x leverage amplifying the losses, banks began issuing margin calls.

This is what Wall Street calls a “margin call”: you borrow money to buy stocks, but when prices drop, banks demand more collateral—or they liquidate your positions. It’s the same logic as forced liquidation in A-shares financing.

SA’s prime brokers included Bank of America, Goldman Sachs, and JPMorgan, all of whom intervened to help the fund handle the margin calls or unwind positions in an orderly fashion.

Of note, Aschenbrenner’s fiancée Avital Balwit is Anthropic CEO Dario Amodei’s chief of staff. The crash happened exactly the weekend before their wedding.

36 Hours, Overnight Auction

The crisis evolved rapidly in 36 hours.

Late on July 29, SA sent urgent requests to multiple potential buyers, seeking to offload over $10 billion in stocks to stabilize the situation. This news was reported widely by the Financial Times, Bloomberg, and others.

Bidders included: Millennium Management (top global multi-strategy hedge fund), Jane Street (an investor in SA).

Millennium made an offer but viewed the portfolio as too risky to pay up.

Ultimately, Citadel prevailed.

By the morning of July 30, the deal was completed. Citadel acquired SA’s entire $16 billion public stock portfolio—including both long and short positions—at a steep discount. It was one of the largest emergency stock transactions in Wall Street history.

Once the deal was reported, the Nasdaq 100 Index jumped over 3% that day—the market breathed a sigh of relief: SA wouldn’t have to dump more stocks onto the market.

This isn’t Citadel’s first time doing such a deal. Previously, it had scooped up companies’ assets like Enron’s at crisis prices. In Ken Griffin’s own words, one reason for Citadel’s continued success is: “Experience. Wisdom bought with loss and pain. My leadership team and I have gone through many of the market’s harshest moments together and learned many painful lessons. But this makes us better investors in times of upheaval and crisis.”

Citadel was founded in 1990. SA was founded in 2024.

What’s Left

After the deal, SA’s assets under management plummeted from a peak of $45 billion to about $10 billion, a drop of more than half.

But Aschenbrenner wasn’t completely wiped out.

SA kept all its private investments, worth about $10 billion, with the most important part being an equity stake in Anthropic, valued at around $5 billion. SA will continue operating as a private investment firm.

Whether these private assets can be realized depends on Anthropic and other companies’ valuations when they go public in the future.

This remains an unsettled question.

By the way: Aschenbrenner’s fiancée Avital Balwit is Anthropic CEO Dario Amodei’s chief of staff. There were rumors that SA would sell Anthropic shares, but an SA spokesperson denied it.

Where Did He Go Wrong?

The most thought-provoking question here is: Was Aschenbrenner wrong in his judgment of AI?

Not necessarily.

The long-term demand logic for AI infrastructure is still widely recognized among institutions.

The problem lay in portfolio structure.

Someone on X quoted Ken Griffin’s description of why fund managers fail: “Your portfolio is highly concentrated with massive positions, but you can’t clearly articulate the competitive edge of holding those positions.”

This was exactly SA’s issue: concentration in a single theme (AI infrastructure), 4x leverage, and bets on both sides of the same narrative. When the market direction changed, there was no buffer at all.

This structure can generate 439% returns on the way up, and can blow up the entire public portfolio in just weeks on the way down.

As summarized by some: “His take on AI wasn’t wrong. He erred in how he constructed his positions.”

The Other Side of the Myth

The Aschenbrenner story is one of the most extreme examples of this round of the AI bull market.

One article, one narrative, eight people, $45 billion, 439% return—and then, within three weeks, leveled by margin calls.

This is neither the end of AI, nor the end of Aschenbrenner.

But it clearly proves one thing: Getting the direction right doesn’t mean you’ll survive until that direction becomes reality.