Leopold Aschenbrenner Predicted AI Perfectly — but Ignored One Brutal Lesson Every Investor Must Learn

Artificial intelligence has rewarded bold investors over the past two years. Companies tied to chips, memory, data centers, and power infrastructure have produced life-changing returns as hyperscalers poured hundreds of billions of dollars into AI. That success also created a dangerous illusion: that a winning investment thesis alone is enough to guarantee investment success. History suggests otherwise. 

Markets have a habit of humbling even the smartest investors, particularly when leverage enters the picture. The rise — and near collapse — of Leopold Aschenbrenner’s Situational Awareness hedge fund is a reminder that being right about the future doesn’t matter if you can’t survive the journey.

A Brilliant AI Thesis Met Wall Street

Aschenbrenner first gained prominence as a researcher on OpenAI’s Superalignment team before leaving the company in 2024 after criticizing what he viewed as inadequate AI security practices. He later published his 165-page manifesto, Situational Awareness, arguing that artificial general intelligence could arrive far sooner than most people expected and that investors should position themselves for an unprecedented buildout of AI infrastructure.

Silicon Valley heavyweights bought into that vision. His hedge fund reportedly launched with roughly $225 million from backers including Stripe co-founders Patrick and John Collison, along with Nat Friedman and Daniel Gross.

The portfolio reflected that conviction.

Rather than buying the biggest technology companies, Aschenbrenner concentrated on businesses he believed would become the “picks and shovels” of the AI revolution. Many still appear well positioned if AI spending continues expanding. The catch is that several are also volatile, speculative stocks whose prices can swing wildly in a matter of weeks.

The strategy looked unstoppable. Reports indicate Situational Awareness generated cumulative gains exceeding 1,000% since launch while assets ballooned beyond $20 billion. Bull markets have a way of making conviction look like genius.

The Market Eventually Tests Every Investor

Warren Buffett famously warned, “Only when the tide goes out do you discover who’s been swimming naked.” That lesson applies just as much to leverage as it does to questionable businesses.

Reports indicate Situational Awareness borrowed aggressively against its portfolio, with leverage reaching roughly four times its equity. A modest amount of leverage can enhance returns when markets cooperate. Four-to-one leverage means a 25% decline can effectively erase an investor’s equity before margin calls force sales. Unfortunately, that’s exactly what happened.

As AI-related stocks corrected, many of Aschenbrenner’s largest holdings fell 30% or more. Rather than waiting for a recovery, margin requirements reportedly forced the fund to liquidate nearly its entire public equity portfolio in a sale to Ken Griffin’s Citadel Advisors. The Financial Times reported the fund lost 67% during July before unwinding much of its public book.

Ironically, many of the companies Aschenbrenner backed may still benefit from the long-term AI boom. His timing wasn’t necessarily wrong. His risk management was.

Key Takeaway

In short, Aschenbrenner’s story isn’t proof that the AI investment theme has failed. It is proof that even the strongest investment thesis can be undone by poor portfolio construction.

Retail investors should take the opposite lesson. If you believe AI infrastructure companies will create enormous value over the next decade, owning a diversified basket of quality businesses without excessive borrowing gives that thesis time to play out. Leveraging a concentrated portfolio may amplify gains during a bull market, but it also removes your ability to survive the inevitable correction.

Ultimately, successful investing isn’t just about predicting the future. It’s about staying in the game long enough to benefit when you’re right.

Contact [email protected] for any questions or corrections.