Ex-OpenAI Researcher Leopold Aschenbrenner Doubles Down on AI Infrastructure and Bitcoin Mining Stocks

Leopold Aschenbrenner, a former researcher at OpenAI, has significantly expanded his public investment portfolio, placing a massive bet on the convergence of artificial intelligence infrastructure and Bitcoin mining. According to filings with the U.S. Securities and Exchange Commission (SEC), his disclosed stock holdings surged from $5.5 billion at the end of last year to $13.67 billion by the end of March this year.

From AI Research to Infrastructure Investment

Aschenbrenner, who previously worked on alignment and safety research at the ChatGPT developer, has pivoted sharply into the capital-intensive side of the AI boom. His filings reveal concentrated positions in companies that operate at the intersection of high-performance computing (HPC) and energy-intensive data centers. These include IREN, Core Scientific, Riot Platforms, CleanSpark, Bitfarms, Bitdeer, and Hive Digital.

The common thread among these firms is their dual capacity to provide power and data center infrastructure for both Bitcoin mining and AI workloads. As demand for AI compute capacity explodes, these companies are increasingly being viewed as critical suppliers of the physical backbone needed to train and run large language models.

A Bet on Infrastructure, A Hedge Against Chipmakers

While Aschenbrenner has accumulated long positions in mining and data center firms, he has simultaneously established significant put option positions against some of the biggest names in the semiconductor industry. According to the filings, he holds put options totaling $7.46 billion against Nvidia, Broadcom, Oracle, and the VanEck Semiconductor ETF (SMH).

This dual strategy suggests a conviction that the value in the AI boom will flow to infrastructure providers and energy operators, rather than solely to chip designers. It reflects a sophisticated view that as AI scales, the bottleneck may shift from silicon to power and data center capacity.

Why This Matters for Investors

Aschenbrenner’s moves are noteworthy not just because of the size of the positions, but because of his insider perspective on the AI industry. Having worked directly on frontier AI models, his investment thesis carries weight. The market has already begun to price in the transition of Bitcoin miners into AI infrastructure providers, with several of these stocks seeing significant volatility and upward movement this year.

For readers, this story underscores a broader trend: the AI industry’s physical infrastructure demands are creating new investment opportunities beyond the obvious chipmakers. It also highlights the increasing financial sophistication of former tech insiders who are translating their domain expertise into concentrated market bets.

Conclusion

Leopold Aschenbrenner’s SEC filings reveal a clear and large-scale investment thesis: the future of AI will be built on infrastructure that Bitcoin miners already operate. By going long on these hybrid companies and hedging against semiconductor giants, he is betting that the next phase of AI growth will be defined by power and data center capacity, not just chip performance. The filings offer a rare, data-backed glimpse into how a former OpenAI researcher is positioning for the AI infrastructure buildout.

FAQs

Q1: Who is Leopold Aschenbrenner?
A: He is a former researcher at OpenAI who worked on AI alignment and safety. He has since transitioned into investment, focusing on AI infrastructure and related technologies.

Q2: Why is he investing in Bitcoin mining companies?
A: Many Bitcoin mining firms own significant power infrastructure, data centers, and high-performance computing capabilities that can be repurposed for AI workloads, making them strategic suppliers in the AI boom.

Q3: What does his put option position against Nvidia mean?
A: It indicates he is hedging against or betting on a decline in Nvidia’s stock price, possibly because he believes the market has overvalued chipmakers relative to the infrastructure providers that will be needed to deploy AI at scale.