Leopold Aschenbrenner is positioning for the next phase of the AI boom by rotating out of semiconductor giants and into bitcoin miners and infrastructure providers that control power and data center capacity.
The former OpenAI researcher—known for his warnings about geopolitical risks to advanced AI—significantly increased his exposure to bitcoin mining and AI infrastructure firms while simultaneously building large bearish positions against chipmakers.
According to his latest 13F filing with the U.S. Securities and Exchange Commission, Aschenbrenner expanded his disclosed equity portfolio from $5.5 billion at the end of 2025 to $13.67 billion as of March 31.
His largest long positions are concentrated in companies tied to energy supply, data centers, and compute infrastructure critical to AI development. These include bitcoin miners and related firms such as IREN, Core Scientific (CORZ), Riot Platforms (RIOT), CleanSpark (CLSK), Bitfarms (BITF), Bitdeer (BTDR), and Hive Digital (HIVE).
Many of these companies have increasingly pivoted beyond crypto mining, positioning themselves as providers of power and high-performance computing capacity for AI workloads. Their access to large-scale energy contracts and existing infrastructure has made them attractive plays as demand for AI data centers accelerates.
The filing also revealed notable positions in Bloom Energy (BE), SanDisk (SNDK), and cloud infrastructure firm CoreWeave (CRWV), underscoring a broader conviction in the long-term growth of AI’s physical backbone.
At the same time, Aschenbrenner has taken an aggressive bearish stance on the semiconductor sector. He disclosed $7.46 billion in put options targeting major chipmakers and related exchange-traded funds, including a $2.04 billion position against the VanEck Semiconductor ETF, a $1.57 billion put on Nvidia (NVDA), and more than $1 billion in combined positions tied to Oracle (ORCL) and Broadcom (AVGO).
The portfolio reflects a clear strategic shift: favoring infrastructure and energy-intensive compute providers over traditional semiconductor leaders in the evolving AI landscape.v
