From Wipeout to Re-Entry in Six Weeks
TREE NEWS reports: Leopold Aschenbrenner, the former OpenAI researcher turned hedge fund manager, is betting on artificial intelligence stocks again — just six weeks after his fund, Situational Awareness, suffered a roughly $35 billion drawdown that nearly sank the vehicle. The fund has purchased call options on Advanced Micro Devices (AMD), Bloom Energy, CoreWeave, SK Hynix, and SanDisk, with the trades executed from late last week into this week.
The move is a striking reversal for a manager whose thesis on AI infrastructure had been punished by a violent rotation out of momentum names. Rather than retreating, Aschenbrenner appears to be treating the selloff as a buying opportunity, using options to express leveraged conviction on the same supply-chain beneficiaries that hurt him.
Why These Names Matter
The basket is not random. It maps directly onto the AI compute stack:
- AMD — the primary challenger to Nvidia in AI accelerators, and a pure-play on inference demand growth.
- SK Hynix — a dominant supplier of high-bandwidth memory (HBM), the bottleneck component in every modern AI GPU.
- CoreWeave — a neocloud GPU operator whose revenue is tied to rented compute contracts.
- Bloom Energy — fuel-cell power provider increasingly pitched as a solution to data-center energy constraints.
- SanDisk — memory and storage exposure as AI workloads drive NAND demand.
Together they represent a full-stack wager: compute silicon, memory bandwidth, rented capacity, and the power to run it all.
Reading the Signal
Two interpretations dominate. The bullish case: Aschenbrenner’s core thesis — that AI capability is compounding faster than markets price, and that infrastructure demand will outrun supply for years — remains intact, and the drawdown simply reset valuations. The bearish case: this is a leveraged re-entry into a crowded trade, and call options amplify downside if the momentum reversal resumes.
What is notable is the instrument. Buying calls rather than stock signals the fund wants convex upside with defined premium at risk — a structure that suggests Aschenbrenner is sizing the position carefully after a near-death experience.
The Crypto Overlap
For crypto markets, the read-through is indirect but real. CoreWeave and similar GPU operators sit adjacent to decentralized compute networks, and any repricing of AI infrastructure demand flows into tokens tied to GPU marketplaces, inference networks, and on-chain AI agents. If Aschenbrenner is right that the AI capex cycle is far from over, the knock-on effect for decentralized compute narratives is constructive. If he is wrong, the same high-beta crypto AI tokens likely sell off harder.
What to Watch
Three things matter from here: whether the options positions are disclosed in the fund’s next filing, how AMD and SK Hynix guide on AI demand in upcoming earnings, and whether power constraints — the Bloom Energy thesis — become the binding limit on data-center buildout. Aschenbrenner has now made his bet twice. The second one will be judged quickly.
