5,000 Still Copy-Trading Aschenbrenner's Fund After Collapse, $28M Remains
Claire Weston
AI hedge fund Situational Awareness lost 67% in July, yet more than 5,000 retail investors on copycat platform Autopilot still hold $28 million in a mirror portfolio — one that strips out the leverage that caused the crash.
The fund nearly blew up — why are people still holding?
Autopilot is a platform that clones publicly disclosed holdings of prominent investors into one-click mirror portfolios for retail traders.
Since Leopold Aschenbrenner's fund launched in March, roughly $32 million flowed into the copycat strategy on Autopilot.
After the July crisis became public, Autopilot sent a memo letting investors choose to stay or exit. Only $3–5 million left. $28 million and over 5,000 investors stayed.
This means → the headline "down 67%" didn't scare most copycats away — because their actual losses were far smaller.
Same person's holdings — why did copycats lose so much less?
The difference comes down to one word: leverage. The fund itself uses leverage, short-sells, and trades complex derivatives. The copycat portfolio does none of that — it only buys stocks.
In plain terms = the fund borrows money to double its bets — wins double, losses double. The copycat portfolio uses only the investor's own cash on the same stocks, so a drop is just a drop, never amplified.
Result: the fund lost 67% in July alone. The copycat portfolio fell far less. From its low point, it has rebounded 35% and is up 53.4% since its March 5 launch.
The flip side: the fund is still up roughly 80% for the year — far above the unleveraged copycat. This reflects the two-edged nature of leverage: it magnifies gains and losses alike.
What are these copycats actually betting on?
The mirror portfolio holds 14 stocks, heavily concentrated in AI infrastructure.
The largest position is emerging cloud company Nebius at about 30%. CoreWeave and Bloom Energy each account for roughly 12%.
These stocks sold off sharply in July on Asian-market volatility and concerns over AI capital spending. They rebounded after Citadel stepped in to acquire most of the fund's public-market positions — a tailwind for copycats who stayed.
Why couldn't copycats act during the crisis?
Autopilot primarily tracks 13F filings — quarterly disclosure reports the SEC requires within 45 days of each quarter's end.
This means → the copycat portfolio updates on a quarterly cycle. Between filings, copycats cannot see or act on any changes the fund makes.
During the July crisis, copycats were effectively in a "forced hold" — they didn't choose to stay; there was simply no new filing to trigger a rebalance.
Aschenbrenner's next 13F is due by August 14. His latest positioning will be closely watched.
What comes next?
Autopilot says it will keep cloning the portfolio as long as Situational Awareness remains operational.
A new variable: if Anthropic completes an IPO later this year, Aschenbrenner's most important private investment enters the public market — and the copycat portfolio's scope.
If the fund ultimately shuts down, Autopilot will handle it case by case. For reference, the platform plans to convert its $535 million Pelosi copycat fund into a general-politician trade tracker after Nancy Pelosi retires early next year.
The core question is simple: the fund has pledged to permanently drop leverage. Whether Aschenbrenner can sustain outperformance without it will determine whether these 5,000-plus copycats ultimately stay or go.
Content is for reference only, not financial advice.