Leopold Aschenbrenner, a former OpenAI employee aged 24, founded the AI-themed hedge fund Situational Awareness, which now manages assets exceeding $20 billion, nearing the scale of Wall Street’s top hedge funds.
Article author and source: AI Era
After leaving OpenAI, 24-year-old AI prodigy Leopold Aschenbrenner is causing a stir.
When he founded the AI-themed hedge fund Situational Awareness two years ago, he had no professional investment experience and started with just a few hundred million dollars.
He now manages assets exceeding $20 billion, approaching the scale of Bill Ackman’s Pershing Square and Dan Loeb’s Third Point—established giants that have been prominent on Wall Street for decades.
As of May this year, the net return after fees is approximately 270%. Since inception, the net gain after fees has exceeded 1000%.
A young person with no investment experience outperformed Wall Street veterans.
Even more striking is the treatment he receives online, with popularity rivaling that of top-tier celebrities.
Although he rarely appears in public or posts online, every quarterly portfolio disclosure from his fund sparks nationwide excitement, with fans analyzing his trades as if studying the Bible, and even apps have launched a “one-click copy trading” feature.
Each market frenzy produces a new “investment guru.”
During the internet bubble, there was Ryan Jacob; during the 2020 bull market, there was “Wood” Cathie Wood.
But when it comes to betting on AI, no one is more celebrated than Leopold Aschenbrenner—yet he barely appears and posts rarely, living like an invisible god.
In May, news emerged that he had invested in the solar energy company T1 Energy, causing the stock price to surge 23% that day, with trading volume reaching the second-highest level in history.
In the popular tech talk show TBPN episode, they spent only two minutes discussing the landmark lawsuit between Musk and OpenAI, but spent more than five times that long talking about his latest holdings.
Podcast host Tim Ferriss dubbed him the “Nostradamus of AI,” saying his predictions about AI have a “near-psychic” accuracy rate.
What exactly did he bet on?
The most profitable trade was Anthropic.
This company now makes up one-fifth of his fund’s assets.
When he invested in February 2025, Anthropic was valued at $61.5 billion; now, that figure is $965 billion.
In less than a year and a half, it multiplied more than tenfold. Such holdings shouldn’t appear in regulatory disclosures—U.S. stock reports don’t include foreign stocks or private companies, and his largest positions are precisely hidden within these two categories.
Also, South Korea’s storage chip giant SK Hynix entered the scene in November 2024 and watched it climb into the trillion-dollar club.
Even the most discerning funds have come to support him.
Wall Street’s most profitable quantitative trading firm, Jane Street, unusually entrusted its funds to him—considering that this firm almost never invests in external fund managers.
Both also co-led the investment in AI chip startup MatX and jointly invested in AI cloud computing company Fluidstack.
But at the beginning of his career, this “Sage of the Stock Market” faced an unsteady journey.
As a German, he graduated from Columbia University in 2021 with the top rank.
He later worked in the charitable division of Sam Bankman-Fried’s FTX cryptocurrency exchange, which later went bankrupt.
He then spent a year at OpenAI, during which there was ongoing controversy, before ultimately being fired.
However, just two months after being fired by the most influential AI giant, he authored a 165-page manifesto on artificial intelligence—Situational Awareness.
Portal: https://situational-awareness.ai/
This manifesto spread like a virus through Silicon Valley and Washington.
Shortly after, he founded a hedge fund managing over $1.5 billion in assets, drawing major tech giants and politicians into his fold.
Each year, the AI safety community produces countless lengthy, obscure, and frustrating academic papers.
Most of them lie quietly on niche forums like LessWrong, fading away amid the debates of a small group of “AI extinctionists.”
But Leopold Aschenbrenner’s June 2024 publication, Situational Awareness, broke this fate.
In this manifesto, Leopold Aschenbrenner avoids the technical jargon that puts investors to sleep.
Instead, he adopted a grand narrative with strong tension.
He attempted to write this manifesto as the AI era’s “long telegram”.
In 1946, American diplomat George Kennan, writing under the pseudonym “X,” authored the “Long Telegram,” inadvertently sparking the Cold War.
Mimicking this scene, Leopold Aschenbrenner is sounding the alarm: the AGI era is approaching at an exponential pace, and the world, aside from a handful of “prophets” in a few labs in Silicon Valley, is utterly unprepared and unable to imagine it.
To help capitalists who know nothing about technology understand the terrifying nature of exponential growth, he made a highly emotionally charged analogy.
At the time, only a very few people truly understood the mathematical implications of the pandemic’s exponential spread.
Before the market completely collapsed, these few individuals shorted the market and hoarded masks, earning historic profits. Today, the arrival of AGI is similarly an irreversible exponential curve.
Those who sense and act early will reap the greatest economic benefits of this century.
This narrative has created a near-religious fervor.
After reading it, University of Texas professor and former OpenAI employee Scott Aaronson exclaimed:
This could be the kind of document that a general or someone from the national security system would read and say, “This requires action.”
He packaged the dull calculations about GPU computing power and data bottlenecks in the lab as a dystopian revelation about human destiny, geopolitical competition, and enormous wealth.
He is not merely predicting the future—he is leveraging people’s intense fear of being left behind (FOMO) to create a safe pathway for capital toward redemption.
Belief is worth nothing in Silicon Valley unless it can be converted into chips.
Sholto Douglas, head of reinforcement learning scaling at Anthropic and a friend of Leopold Aschenbrenner, described it as a “theory of change”:
Leopold Aschenbrenner believes how the world will evolve, and he is indeed putting his own funds into the places he speaks about.
Thus, Situational Awareness LP hedge fund was established.
The fund is seeded by Silicon Valley heavyweights including Nat Friedman, head of Meta AI products, Daniel Gros, and the Collison brothers, co-founders of Stripe, with Carl Shulman, an AI forecasting scholar, serving as research director.
The logic of the SA fund is extremely straightforward:
If superintelligence is destined to arrive, then regardless of which AI giant ultimately prevails, humanity must build temples for it in the physical world—
This requires astronomical amounts of electricity, transformers, copper cables, and semiconductors.
According to publicly disclosed holdings filed with the U.S. Securities and Exchange Commission (SEC), SA Fund’s long positions are precisely aligned with these AI physical infrastructure assets:
- Intel and Broadcom: The chip foundation enabling connectivity for trillion-parameter models.
- Vistra: A leading independent power producer providing uninterrupted energy to data centers.
- Core Scientific: A rising infrastructure provider that successfully transitioned from a Bitcoin mining company to a leader in high-density AI computing hosting.
This “unity of knowledge and action” hands-on investment strategy generated an impressive 47% net return for the SA Fund in the first half of 2025.
He remarkably allocated approximately $459 million in Intel call options in the first quarter, accurately predicting the subsequent U.S. government capital injection and NVIDIA’s equity stake, leaving even seasoned Wall Street veterans stunned.
Senior hedge fund investor Graham Duncan remarked that Leopold Aschenbrenner reminded him of Michael Burry from The Big Short: “If you want to have a differentiated view, you’d better be a bit different yourself.”
However, behind the glamorous wealth myth of Leopold Aschenbrenner lies a deeply dark and unsettling moral shadow.
Do you also feel the absurdity and fragmentation here?
Leopold Aschenbrenner was a core member of the Effective Altruism (EA) community, whose ultimate mission is to prevent existential risks posed by AI that could lead to human extinction.
He previously worked on OpenAI’s Superalignment team, studying daily how to ensure machines smarter than humans remain kind.
His fiancée, Avital Balwit, currently Chief of Staff to the CEO of Anthropic, has also written an article expressing concern that AGI will end human employment within five years.
However, while his fiancée was worrying about “the end of work,” Leopold Aschenbrenner was using the same technology to make his investors wealthy.
For former colleagues in the security community, Leopold Aschenbrenner’s actions amounted to the most complete betrayal.
“Some people who are deeply concerned about AI existential risks now strongly dislike Leopold. They feel he sold out his ideals,” lamented a former OpenAI governance researcher.
The deeper crisis lies in the narrative of “geopolitical arms race” he promotes.
In “Situational Awareness,” he strongly advocates that the United States must accelerate AI development at wartime mobilization speed, at any cost, to maintain global technological dominance.
This dangerous rhetoric plays into the hands of Washington hawks and Silicon Valley accelerationists and is beginning to genuinely influence U.S. national strategy.
As AI governance experts have warned: “The best way to predict a technological cold war is to fund it yourself.”
By instilling in politicians and capitalists the fear that “if we don’t accelerate, we’ll be destroyed,” Leopold Aschenbrenner personally propelled the arrival of the very “dangerous future” he described.
Meanwhile, his hedge fund profited from the arms race he himself incited.
He successfully proved that narrative is wealth.
On the eve of AGI, the most profitable are not those who write code, but the super marketers who can translate “tech panic” into “100x returns.”