Leopold Aschenbrenner, a former OpenAI employee, is putting $13.6 billion behind an AI investment strategy that flips the conventional playbook on its head. He's shorting Nvidia and AMD while making a big bet on bitcoin miners — because they control the electricity and data centers he sees as critical for the next AI phase.
Shorting the chip giants
While the broader market has cheered Nvidia and AMD as the undisputed winners of the AI boom, Aschenbrenner sees overvaluation. His short positions target both companies directly. It's a bold call, especially given the rally in AI stocks over the past two years. But he's not alone in questioning whether the chip makers' growth can keep up with their stock prices.
Why miners — not just miners
Bitcoin miners have long been dismissed as crypto plays. Aschenbrenner sees them differently. These firms own power purchase agreements, substations, and massive cooling systems. That infrastructure is exactly what AI training clusters need. The thesis: as AI models scale, the fight won't just be for GPUs — it'll be for megawatts. Miners already have those megawatts locked in.
A $13.6 billion conviction
The scale of the bet is hard to ignore. Aschenbrenner isn't making a small allocation — he's reportedly executing a strategy worth over thirteen billion dollars. That kind of money can move markets, especially in the relatively small miner sector. It also means he's deeply exposed if his timing is off or if the miners fail to pivot their business models fast enough.
The open question
Can bitcoin miners actually become AI infrastructure providers? A few, like Core Scientific and Hut 8, have already signed deals with AI companies. But the transition isn't automatic. It requires changes in hardware, software, and customer relationships. Aschenbrenner is betting the market will eventually price miners as hybrid energy-AI assets. Whether that happens in time to justify his shorts on Nvidia and AMD remains the big question.
