A former OpenAI employee turned fund manager is betting against the very stocks that build it

Leopold Aschenbrenner, a former OpenAI researcher fired in April 2024, has taken large short positions worth over $7 billion against AI and semiconductor stocks like Nvidia, AMD, and Oracle through his hedge fund, Situational Awareness LP. His fund also holds long positions in energy and Bitcoin mining firms, reflecting his belief that AI advancement will depend on physical infrastructure like electricity and computing power." "article": "Leopold Aschenbrenner, a former OpenAI researcher and Columbia University valedictorian, has made a controversial move in the financial markets. His hedge fund, Situational Awareness LP, recently disclosed over $7 billion in put-option positions targeting major AI and semiconductor companies, including Nvidia, AMD, Oracle, and Broadcom. The largest short position is a $2.04 billion bet against the VanEck Semiconductor ETF, a fund tracking semiconductor stocks. Aschenbrenner joined OpenAI’s Superalignment team in 2023, focusing on ensuring AI systems remain safe and aligned with human interests. However, his tenure ended abruptly in April 2024 when he was fired over an alleged information leak, a claim he disputes. He argues his dismissal was retaliation for an internal memo warning of security risks and potential foreign industrial espionage at OpenAI. After leaving OpenAI, Aschenbrenner published a 165-page essay titled *Situational Awareness: The Decade Ahead*, predicting that Artificial General Intelligence (AGI) could emerge as early as 2027. Despite his bearish stance on AI-related stocks, his fund holds significant long positions in energy-intensive infrastructure, including Bloom Energy ($878 million) and Bitcoin mining firms like IREN, Core Scientific, and CoreWeave. This strategy reflects his view that AI’s growth will be constrained by physical infrastructure, particularly electricity and computing power. As of March 31, Situational Awareness LP’s equity exposure surged to $13.6 billion, more than doubling from $5.5 billion at the end of December 2025. The fund’s aggressive betting against AI and semiconductor giants contrasts with its bullish bets on energy and mining, underscoring Aschenbrenner’s conviction that infrastructure bottlenecks will shape the future of AI development.
Leopold Aschenbrenner, a former OpenAI researcher and Columbia University valedictorian, has made a controversial move in the financial markets. His hedge fund, Situational Awareness LP, recently disclosed over $7 billion in put-option positions targeting major AI and semiconductor companies, including Nvidia, AMD, Oracle, and Broadcom. The largest short position is a $2.04 billion bet against the VanEck Semiconductor ETF, a fund tracking semiconductor stocks. Aschenbrenner joined OpenAI’s Superalignment team in 2023, focusing on ensuring AI systems remain safe and aligned with human interests. However, his tenure ended abruptly in April 2024 when he was fired over an alleged information leak, a claim he disputes. He argues his dismissal was retaliation for an internal memo warning of security risks and potential foreign industrial espionage at OpenAI. After leaving OpenAI, Aschenbrenner published a 165-page essay titled *Situational Awareness: The Decade Ahead*, predicting that Artificial General Intelligence (AGI) could emerge as early as 2027. Despite his bearish stance on AI-related stocks, his fund holds significant long positions in energy-intensive infrastructure, including Bloom Energy ($878 million) and Bitcoin mining firms like IREN, Core Scientific, and CoreWeave. This strategy reflects his view that AI’s growth will be constrained by physical infrastructure, particularly electricity and computing power. As of March 31, Situational Awareness LP’s equity exposure surged to $13.6 billion, more than doubling from $5.5 billion at the end of December 2025. The fund’s aggressive betting against AI and semiconductor giants contrasts with its bullish bets on energy and mining, underscoring Aschenbrenner’s conviction that infrastructure bottlenecks will shape the future of AI development.
This content was automatically generated and/or translated by AI. It may contain inaccuracies. Please refer to the original sources for verification.