The "madman" fired by OpenAI earned 5.5 billion USD in 2 years

Today, let’s talk about the recently popular “child prodigy + controversial figure + new AI investment star” in Silicon Valley’s AI circle—Leopold Aschenbrenner. Many of you should have come across him in various reports.

Mainly, this guy is only 24 years old, and the fund he manages has reached 5.5 billion dollars. When it was founded less than two years ago in 2024, it was only 250 million, which has increased twentyfold in under two years.

There was also a previous viral incident: he was a member of OpenAI’s “Superalignment” team, but after being expelled, he wrote a groundbreaking AI article that shook Silicon Valley—“Situational Awareness,” which has now been read over 5 million times.

Many now see him as: “The early AI version of Cathie Wood + Peter Thiel + a tech prophet.”

1. His background: The super-elite academic route

Leopold was born in Germany, with both parents being doctors. A very typical European elite education path. From a young age, he demonstrated strong math and scientific research abilities: winning awards in German youth science competitions, then attending Columbia University in New York to study economics and mathematical statistics. At 19, he was the valedictorian at Columbia, a classic “child prodigy”!

He was introduced early to the Effective Altruism (EA) community. This circle later became deeply connected with AI and cryptocurrency, such as FTX’s founder, Ilya Sutskever (Chief Scientist), who later led him into OpenAI, and many people from Anthropic and OpenAI.

2. His “rise to wealth”: From OpenAI to AI hedge fund

1. Joining OpenAI’s Superalignment team

In 2023, he joined OpenAI, specifically the core “Superalignment” team, led by Ilya Sutskever and Jan Leike.

This team mainly researches how to control AI if it surpasses human intelligence—one of OpenAI’s most sensitive and core research areas.

2. His expulsion from OpenAI

In 2024, a major turning point occurred. He was expelled from OpenAI over “information leakage.” However, Leopold claims he saw security issues at OpenAI, wrote a security memo to the board, but management became very nervous and fired him. The entire Superalignment team was also disbanded. Many say Leopold’s departure was part of an internal power struggle within OpenAI.

3. The real reason he became a legend: “Situational Awareness”

After being expelled, he immediately wrote an article: “Situational Awareness: The Decade Ahead,” which had a huge impact in Silicon Valley’s AI circles.

Its core ideas are very radical. He predicts:

1. AGI will emerge around 2027

He believes AI development is not linear but a “combustion cycle of compute power + data + automated AI research.”

By around 2027, AI might be capable of conducting AI research independently.

2. Then, an “intelligence explosion”

He thinks: if AI begins to autonomously develop AI, it could compress 10 years of research into 1 year, rapidly surpassing humans!

This is similar to “recursive self-improvement,” continuously accelerating 24/7, which is terrifying.

3. The real bottleneck isn’t models but infrastructure

This view is similar to Huang Renxun’s earlier articles, emphasizing that current AI limitations are power supply, data centers, chips, fiber optics, and energy. This perspective later became part of his investment logic.

4. His investment approach: AI Infrastructure Thesis

After leaving OpenAI, he founded Situational Awareness LP, which specifically invests in AI-related industries. Backers are top-tier Silicon Valley AI/tech capital circles.

5. His investment track record

Founded in 2024 with a 250 million dollar fund, reaching 1.5 billion in 2025, and now 5.5 billion. For someone in their early 20s without traditional Wall Street experience, this is extraordinary.

Returns are also impressive. From his portfolio, the largest holding is Bloom Energy, which mainly supplies power to data centers—estimated to have tripled since 2024. He also holds options on CoreWeave and Intel—CoreWeave is an AI cloud service provider, a typical “AI pick-and-shovel” company; Intel, needless to say, has also multiplied many times. Options amplified his gains significantly.

Additionally, he invested in SanDisk, which has increased 40-fold in two years. His investment strategy aligns with his articles’ ideas. His portfolio also includes Cipher Mining, a Bitcoin mining company, but he isn’t bullish on Bitcoin itself. Instead, he sees mining farms, computing power, and machines as assets that can be directly transformed into AI data centers. This company has also signed large leasing contracts with AWS and Google for computing capacity.

6. Controversies

Of course, he is not without controversy. Many criticize his views as too extreme—for example, he emphasizes long-term AI national competition, US-China AGI rivalry, and claims that in the next decade, the US must not fall behind China in AI.

Some also believe that the 2027 AGI replacing humans is too soon, and many are skeptical.

Currently, the crypto industry is paying attention to him because many of his ideas overlap with crypto. For example, AI needs energy, decentralized computing power, data centers, Bitcoin miners transitioning to AI infrastructure, and re-pricing of power assets.

Now, many Bitcoin mining companies, AI + crypto projects, and DePIN projects cite Leopold’s logic.

7. Brief summary:

Leopold is less like a traditional fund manager and more like an AI technology trend prophet or a builder of the AI worldview. He sees this AI revolution as an “industrial revolution,” not just like the previous internet revolution.

Therefore, updating our understanding is crucial to survive and profit in the market!

Because the most profitable times are often when: old logic fails, and new logic is just beginning.