AI hedge fund Situational Awareness may have sold its public portfolio, but it still has its Anthropic shares

We have all seen this movie before. A brilliant researcher leaves a top-tier lab, writes a manifesto that goes viral, and launches a fund to capitalize on the coming technological singularity. Leopold Aschenbrenner, a former OpenAI researcher who was famously ousted for concerns regarding security and leakers, is currently living through the sequel: the reality check.

The Leveraged Bet Gone Wrong

Recent reports indicate that Aschenbrenner’s fund, Situational Awareness, has effectively liquidated its public equity portfolio. For a fund that was supposed to be the vanguard of AI-driven investment, this is a sobering moment. The fund wasn't just buying stocks; it was using significant leverage to bet on the continued explosion of public AI companies. When the market corrected, that leverage turned into a trap.

As a founder, there is a lesson here about the difference between being right and being timed right. You can be 100% correct about the direction of a technology and still get wiped out by the volatility of the financial instruments you choose. Aschenbrenner’s thesis is that we are on a fast track to Artificial General Intelligence (AGI) and that the economic implications will be total. That might be true. But the public market doesn't trade on ten-year geological shifts; it trades on quarterly earnings and interest rate whispers.

The Anthropic Life Raft

Despite the fire sale in the public markets, Situational Awareness isn't dead yet. The fund still holds its private shares in Anthropic. This is the ultimate hedge. While the public portfolio was likely designed to provide liquid gains, the Anthropic stake represents the core of the fund's actual value proposition: access to the inner sanctum of the AI labs.

Anthropic remains the primary rival to OpenAI, and its valuation continues to defy the gravity that hit the public tech sector. For Aschenbrenner, holding onto these shares is a survival mechanism. It allows him to remain relevant in the cap table wars, even if his public trading desk is currently quiet. It’s a classic move: lose the battle in the streets (public markets) to keep the fortress (private equity).

The Founder Perspective: Signal vs. Noise

For those of us building in this space, the fall of a high-profile AI fund is a signal to ignore the noise of the "AI bubble" talk. The fact that a fund manager made a bad leveraged bet on public stocks doesn't mean the underlying technology is failing. It means the financialization of AI has reached a peak of irrationality.

We are seeing a massive disconnect between the builders and the betters. The builders are still grinding on latency, context windows, and agentic workflows. The betters are trying to turn those engineering breakthroughs into 10x leverage plays. When the leverage breaks, the skeptics come out to say AI is over. Don't believe them. The liquidation of a public portfolio is a math problem, not a technology problem.

Why Public Markets are Treacherous for AI Purists

• Volatility: AI stocks are currently the primary vehicle for macro speculation, making them decoupled from actual product utility.
• Timing: AGI might be coming in 2029, but a margin call can happen on a Tuesday in 2026.
• Liquidity: Public markets provide an exit, but they also provide a way to lose everything faster than a private company valuation can drop.

The Security Researcher’s Paradox

There is an irony in Aschenbrenner’s situation. His entire brand is built on "situational awareness"—the idea that we must be hyper-vigilant about the risks and trajectories of AI. Yet, he seemingly lacked awareness of the basic market risks associated with leveraged trading. It’s a reminder that being a genius in one domain (AI safety and policy) does not automatically make you a genius in another (macroeconomic trading).

This is a common trap for founders. We get so deep into our own vertical that we assume we can apply that same logic to everything else. If you can understand transformer architecture, you think you can understand the bond market. The market, however, is a different kind of intelligence—one that is often more chaotic and less logical than any neural network.

What Builders Should Take Away

If you are building an AI startup right now, look at the Situational Awareness story as a cautionary tale about focus. Aschenbrenner tried to be a thought leader, a policy advocate, and a hedge fund manager all at once. The result was a public setback that overshadowed his intellectual contributions.

Keep your cap table clean and your bets focused on what you can control. The public markets are a casino; your product is the only thing that is real.

We are entering a phase where the "easy money" in AI is gone. The period where you could just say "AI" and get a check is over. Now, we are in the deployment phase. The funds that survive will be the ones that hold stakes in the fundamental infrastructure—like Anthropic—rather than those trying to day-trade the hype cycle.

The Path Forward

Don't expect Aschenbrenner to disappear. The Anthropic stake gives him a seat at the table for years to come. But expect the rhetoric to shift. The "manifesto" era of AI is ending, and the "utility" era is beginning. The markets are forcing a pivot from grand theories to actual balance sheets.

For the rest of us, the message is clear: stay grounded. If the smartest guys from OpenAI can get their portfolios liquidated, you aren't immune to the cycle. Build something that people pay for today, so you don't have to rely on a leveraged bet on tomorrow.

Read the original at TechCrunch AI →