Hook
Did you catch the 13F filing from Situational Awareness LP on August 15? On the surface, it looks like a genius bet on the AI compute boom. But dig deeper, and you'll see a portfolio that screams vulnerability. Between Q1 and Q2 2026, Leopold Aschenbrenner's fund went from a mixed long-short strategy to a near-all-in long on AI hardware and infrastructure. Micron and SanDisk alone now account for over 55% of the public stock portfolio. That's not conviction—it's a time bomb waiting for a sell-off.
Context
Situational Awareness LP is managed by Leopold Aschenbrenner, a former OpenAI researcher known for his apocalyptic views on AI risk. His fund's 13F filings are a window into how the smart money views the AI infrastructure buildout. In Q1, the fund held significant put options against SMH, NVIDIA, Broadcom, AMD, Oracle, Micron, and TSMC—a classic hedge against a tech correction. But by June 30, those puts were slashed to near-zero. The fund now holds a concentrated long position in storage chips, AI compute, and energy infrastructure. The shift is dramatic: Micron holdings surged from $5.86 million to $5.574 billion; SanDisk from $724 million to $5.674 billion. Bloom Energy, TSMC ADR, Nebius, CoreWeave, Core Scientific, Applied Digital, IREN, Riot—all became heavy holdings. This is not a diversified portfolio. It's a bet on a single narrative: that AI compute demand will grow unabated, forever.
Core
Let's break down the numbers because numbers don't lie—they scar. At the end of Q2, Micron and SanDisk together represented ~55% of the fund's disclosed equity portfolio. That's roughly $11.25 billion in two storage chip stocks. Add in the other AI infrastructure names—Bloom Energy ($1.9B), TSMC ($1.26B), Nebius ($1.23B), CoreWeave, Core Scientific, Applied Digital, IREN, and Riot—and you get a portfolio almost entirely correlated to the AI hardware cycle. The risk is not just concentration; it's correlation. If storage demand falters, or if hyperscalers cut their capex, every position gets hit simultaneously.
Based on my audit experience during the 2017 Ethereum mania, I learned that when a portfolio loses its hedging mechanism, it's only a matter of time before a small crack becomes a chasm. In Q1, Situational Awareness held puts on the very stocks they now own. That was a disciplined strategy: long the story, short the volatility. By Q2, they removed the safety net. The fund now relies entirely on the market continuing to believe in the AI compute narrative. But the market is a fickle beast.
Since July, the AI chip and storage sector has faced multiple sell-offs. Micron, SanDisk, SK Hynix all dropped. The Philadelphia Semiconductor Index recorded a rare monthly decline. Then, in early August, cooling inflation data and a recovery in AI earnings sentiment sparked a rebound. SanDisk, Micron, CoreWeave, Nebius bounced hard. But the rebound doesn't erase the risk. The portfolio is still sitting on a knife's edge: any macro shock or narrative shift could trigger a chain reaction. If leverage or margin calls are in play, a simple sector pullback could spiral into a liquidity crisis.
Contrarian
The conventional wisdom says: Aschenbrenner is a visionary, he's betting on the future of AI, and the recent rebound proves his thesis. But I see a different story.
Every scar in the market teaches a new rule. The 2022 Terra Luna collapse taught me that concentrated bets, even on seemingly sound narratives, can destroy trust in hours. When I faced my own community after Luna, I had to rebuild trust by admitting my models were flawed. Aschenbrenner may be a genius, but his portfolio is a textbook example of what happens when conviction replaces risk management. The lack of hedges is not confidence—it's arrogance. The market doesn't care about your vision; it cares about liquidity.
We walk away from greed, we stay for trust. Trust is the only asset that survives the crash. In this case, the fund is trusting the AI narrative to hold, but it's not trusting its own ability to survive a 20% drawdown. The high correlation of positions means that the fund's entire net worth is tied to the same set of factors: storage demand, electricity prices, and hyperscaler spending. That's a fragile foundation.
Takeaway
What does this mean for you? Whether you're trading AI stocks or crypto, the lesson is the same: concentration kills. Leopold Aschenbrenner's 13F is a warning, not a blueprint. The rebound in August is a temporary reprieve, not a vindication. If the market turns, these positions will bleed together. The real question is: how many traders are copying this strategy without understanding the risk? And when the next crash comes, will the market have the liquidity to absorb the cascade?
Trust is built on transparency, not leverage. Protect the flock, not just the profits.