This is a cross-post from Polymarket Institutional Research. To receive future issues, sign up here.
The Polymarket Institutional Desk offers block trades on U.S. and international exchanges, institutional market maker relationships, and enterprise data services. Contact Brooke Rizzetto at institutional@polymarket.com for onboarding.
On July 30, margin calls forced AI hedge fund Situational Awareness to sell its $16 billion public-equity book to Citadel after the fund cruised into July’s AI selloff at a reported 4X leverage.
This article examines how this episode played out on Polymarket, in particular the “AI bubble burst in 2026” contract, which is a near-perfect proxy for the market forces that no doubt made Leopold’s pre-wedding investment “colloquium” far more stimulating.
But were Polymarket traders simply responding to the news headlines? Or anticipating them?
Drawdown gauge
Polymarket’s AI bubble burst contract resolves YES if at least three of six conditions occur within a 90-day window before year-end.
The criteria are a mixture of equity drawdowns, compute pricing, and AI lab corporate events (eg bankruptcy) that would result from significant market pressures:
These are the exact moves that would also push a highly leveraged AI fund like Situational Awareness to its breaking point.
By comparing the movement of this AI stress index to the news flow around Leopold’s fund, we see that this Polymarket contract anticipated many of the key developments in the saga.
The above chart plots the Polymarket AI burst odds alongside the VanEck Semiconductor ETF, a widely cited proxy for the AI trade.
Notice that the Polymarket contract didn't move higher until the Situational Awareness stress emerged in late July, while the VanEck Semiconductor ETF had been falling all month.
In this instance, the Polymarket contract arguably reflected the acute stress around the fund unwind more directly than one of the most-watched AI proxy ETFs.
What moved first?
Prior to the Financial Times announcement on the evening of July 29 that Situational Awareness was seeking capital, Polymarket traders had already anticipated that the fund was in trouble.
For the two weeks leading up to the margin call, the market sat in a narrow 15-to-17% range, even as the semiconductor index was already rolling over. Then, on July 27, traders broke this range, bidding the odds to 20%. By the afternoon of July 29 the market had repriced to 22%; the FT’s report that evening only took it the rest of the way, to its 28% peak.
Right now, the AI bubble burst market is priced at 12%, aligning more with its pre-Situational odds.
Another market also seemed to anticipate this event. NVIDIA's odds on Polymarket of being the world's largest company at the end of August fell from 78% on July 24 to 36% on July 30. This anticipated the dip in the AI hardware equities that Leopold was so highly leveraged on.
Containment or contagion?
Within a week, Polymarket’s AI stress contracts had round-tripped the entire move and kept going. By August 6, the AI bubble market sat near 15%, below the range it held before the margin calls. The largest AI-fund blow-up of the cycle just happened, and the priced odds of an AI bubble burst are lower than before. Instead of treating this sale as a forewarning about the future of AI, Polymarket traders seem to be pricing this event as a risk isolated to a single overextended fund.
NVIDIA’s end-August dominance odds, which had cratered to 36% on the day of the sale, are back above 93%, higher than before the unwind. Even the broad US recession contract barely twitched, holding at 10–12% throughout. The polymarket on whether the fund itself winds down is priced at just 5% by year-end.
On August 5, Situational Awareness announced it was closing a $400 million private investment in a Sequoia-backed company, on top of $100 million it had put into the same company a month earlier.
Polymarket traders reacted early
Across a handful of Polymarket contracts, you could watch as systemic risks to the AI trade were repriced in real time. Equities markets repriced the AI complex in the days before the margin call, jolted when the sale broke, and have climbed steadily back since, capped by the fund’s own return to investing. But while equities gyrated wildly as traders hunted Leopold’s troubled trades, Polymarket moves were more modest, never pricing the AI bust scenario above 30%.
For both retail traders and institutional desks, this case study shows the value of looking at event contracts as a contra-indicator to index proxies to sort the signal from the noise during periods of market dislocation.
Related Polymarkets
Disclaimer
Nothing in The Oracle is financial, investment, legal or any other type of professional advice. Anything provided in any newsletter is for informational purposes only and is not meant to be an endorsement of any type of activity or any particular market or product. Terms of Service on polymarket.com prohibit US persons and persons from certain other jurisdictions from using Polymarket to trade, although data and information is viewable globally.
the real market to watch — the burst probability spiking to 25%+ exactly when the fund’s forced liquidation was happening isn’t a coincidence, that’s the market pricing in a single overleveraged position as if it were systemic evidence, then correcting once the seller was actually just… one seller.