Ex-OpenAI's Leopold Aschenbrenner Bets Big on Crypto Miners for His $13.6 Billion AI Play

Why It Matters

The strategy underscores a growing belief that AI’s next growth phase will be powered by existing mining assets, potentially reshaping capital flows away from chip makers toward energy‑intensive infrastructure providers.

Key Takeaways

• Aschenbrenner raised AI‑related equity to $13.7 billion.
• Long bets focus on bitcoin miners supplying AI power.
• Short positions total $7.5 billion against Nvidia and chip ETFs.
• Miner firms pivot to AI data‑center services.
• Strategy bets on energy‑rich infrastructure over traditional chips.

Pulse Analysis

Leopold Aschenbrenner, once a researcher at OpenAI, has become a high‑profile contrarian investor by dramatically increasing his AI‑related holdings to about $13.7 billion. His latest 13F filing reveals a portfolio heavily weighted toward bitcoin miners such as Riot Platforms, Core Scientific, and Bitfarms, as well as ancillary firms like Bloom Energy and CoreWeave that supply power and compute capacity. By targeting companies that already own large‑scale electricity contracts and data‑center footprints, Aschenbrenner is betting that these assets can be repurposed to meet the soaring demand for AI training and inference workloads.

The shift reflects a broader industry trend where crypto‑mining operations are repositioning themselves as AI infrastructure providers. Their existing high‑density power usage, cooling systems, and proximity to cheap energy sources give them a cost advantage over building new data centers from scratch. Investors are watching how miners negotiate power‑purchase agreements and expand into AI‑specific hardware, potentially creating a new class of hybrid facilities that serve both blockchain and machine‑learning workloads. This convergence could accelerate the deployment of AI services while offering miners a diversified revenue stream beyond volatile cryptocurrency prices.

At the same time, Aschenbrenner has taken roughly $7.5 billion in bearish bets against semiconductor giants, including a $1.57 billion put on Nvidia and a $2.04 billion put on the VanEck Semiconductor ETF. The contrast highlights his skepticism that traditional chip makers will capture the full upside of AI growth, especially if power‑rich mining sites become the primary compute backbone. Market participants should monitor the performance of these miner‑centric stocks and the evolving regulatory landscape, as any shift in energy policy or AI compute demand could dramatically impact the balance between chip‑centric and infrastructure‑centric investment theses.