Michael Burry: The Big Short, the AI Bubble Bet, and the $1.1 Billion That Was Really $9 Million
Michael Burry is the rare trader famous enough to be played by a movie star. In The Big Short, he’s the socially awkward, heavy-metal-blasting fund manager who read the actual mortgage documents nobody else bothered with, concluded the American housing market was a bubble, and bet against it years before it broke — sitting through withering losses and a near-revolt from his own investors until, in 2007 and 2008, he was proven overwhelmingly right. It is one of the great contrarian trades in market history, and it made him a permanent icon of the lone analyst who sees what the crowd cannot.
Which is exactly why his most recent move is so instructive — not because of what he bet, but because of what the coverage of it reveals about how markets, and traders, misread information.
Key takeaways
- Burry made his name betting against subprime before 2008 — the deep-value, sit-in-the-pain contrarian at the center of The Big Short.
- In late 2025 his Scion fund disclosed puts on Nvidia and Palantir, reported everywhere as a ~$1.1 billion bet against AI.
- But that’s the notional value. Burry said he spent about $9.2 million on premium — and for a put buyer, the premium is the whole risk. The headline was 120× the stake.
- The transferable lessons are conviction paired with defined risk, and the brutal truth that being early is indistinguishable from being wrong — which is why sizing, not the call, decides survival.
The trade that made him: sitting in the pain
The Big Short is worth revisiting for one detail that the film gets right and most retellings underplay: Burry was correct and in agony for a long time. He began buying credit default swaps on subprime mortgage bonds in 2005 — effectively insurance that paid off if the bonds failed. For nearly two years the bonds didn’t fail, the premiums bled his fund, and his investors were furious enough that some tried to pull their money and sued to get out. He reportedly restricted redemptions to hold the position. Then the mortgages went bad, the swaps exploded in value, and the trade paid off on a scale that entered legend — reportedly hundreds of millions for his investors and roughly $100 million for Burry personally.
Notice what that story is actually about. It is not “genius spots bubble.” It is “correct thesis plus the structural ability to survive being early.” Burry’s edge was analytical, but his payout depended entirely on a risk structure that let him stay in the trade through the years when he looked wrong. Change that one variable — give him a position that forced him out during the drawdown — and the same correct analysis produces a catastrophic loss. Hold that thought, because it is the entire key to reading his latest bet.
The AI bet, and the number everyone got wrong
In November 2025, Scion Asset Management’s quarterly disclosure lit up financial media: Burry, the man who shorted the last great bubble, had put options on Palantir (on five million shares) and Nvidia (on one million shares), the two totems of the AI trade. The figure that ran in the headlines was roughly $1.1 billion — a billion-dollar bet against artificial intelligence from the Big Short guy himself. It was irresistible, and it was misleading.
Because that $1.1 billion is the notional value — the market value of the underlying shares the options reference — not the money Burry put at risk. When you buy a put, your maximum loss is the premium you paid, full stop. And Burry, to his credit, said so plainly on X: he’d spent about $9.2 million buying roughly 50,000 Palantir puts, with the right to sell at $50 into 2027. The real stake was on the order of one percent of the number the world was quoting.
The headline said $1.1 billion. The trader said $9.2 million. The gap between those two numbers is one of the most useful lessons in finance — and it costs nothing to learn.
This is not a footnote; it is the whole point for a retail audience. Filings are routinely reported in ways that make a modest, defined-risk position look like a leviathan. A trader who reads “Burry bets $1.1 billion against Nvidia” and adjusts their own behavior — piling into a short because a legend supposedly has — has been misled twice over: once about the size, and once about the fact that they can see the position but not the sizing, the timeline, or the reasoning behind it. As our review of Fooled by Randomness argues, and our essay on attention-driven buying measures, the loud number is exactly the one to distrust.
The Cassandra problem: right, early, and disbelieved
Around the same time, Burry did something even more telling. In October 2025 he told investors he was winding Scion down — his read on value, he said, was “not now, and has not been for some time, in sync with the markets” — and by mid-November the fund was deregistered with the SEC. He then re-emerged with a paid newsletter, pointedly titled Cassandra Unchained, laying out the case that the AI boom is a bubble.
The name is the tell. Cassandra, in the myth, was cursed to prophesy truly and be believed by no one. Burry has embraced the label for years, and it captures the contrarian’s central torment: being right about what is not the same as being right about when, and the market can punish a correct thesis for longer than most capital can endure. Burry has been early and loud before — bearish calls that the market ran straight through, positions closed before they would have paid. Being early is not a milder version of being right. In real time, with real money and real drawdowns, being early is identical to being wrong, right up until the instant it isn’t.
This is the honest counterweight to the legend, and a party-neutral profile has to state it: we are not endorsing Burry’s AI view, and history offers plenty of examples of bubbles that ran years past the first credible warning. The value of studying him is not the call. It’s the structure around the call.
The lesson isn’t “short AI.” It’s the opposite of a tip. Burry’s own behavior — capping his risk at a small premium, expressing a multi-year thesis through defined-risk options rather than a naked short — is a study in how to hold a contrarian view without letting it bankrupt you if you’re early. Copy the risk structure, not the position.
What’s actually transferable
Strip away the celebrity and Burry leaves a retail trader three genuinely useful habits, none of which is a stock pick.
First, do your own primary work: his subprime edge came from reading the loan-level documents the crowd assumed someone else had checked. Second, define your risk before you need to: a put buyer knows the worst case on day one, which is what lets a contrarian survive being early — the single variable that made the Big Short pay. Third, separate conviction from sizing: Burry can be maximally convinced and still risk only a small, capped premium, because conviction is a reason to take a position, never a reason to bet the account. That last one is where most retail traders, flush with a strong opinion, do the real damage — and it’s the one Burry models best.
From belief to behavior: read Burry’s discipline in your own trades
Each of Burry’s actual lessons — as opposed to his positions — is measurable in your own record.
| The Burry lesson | The fingerprint it leaves in your trade history |
|---|---|
Conviction ≠ sizing | Position size that stays controlled even on your highest-conviction ideas. If size balloons with confidence, that’s the size discipline leak that being-early turns fatal. |
Defined risk | A capped, knowable worst case per trade; a worst- loss-to-typical-win ratio that no single position can blow open. |
Don’t copy the headline | Entries triggered by news of what a famous investor “is doing” rather than your own setup — attention trading, and usually late. |
Surviving being early | Are your losers cut to a plan, or held with growing size in the hope of vindication? The second is where a good thesis becomes a blown account; see after-loss tilt. |
Burry’s edge was never just the call — it was risking a small, defined amount on it. Upload a broker statement and Gecko scores your sizing, tilt, and overtrading in dollars across twelve behavioral axes, so you can see whether your conviction trades are controlled or quietly betting the account. No login or broker connection needed, first 100 trades free.
An educational tool, not financial advice.
Resources and further reading
- The origin story: Lewis, M. (2010), The Big Short: Inside the Doomsday Machine — the definitive account of Burry’s subprime trade and the investors who doubted him.
- The 2025 disclosure: reporting on Scion Asset Management’s Q3 2025 13F puts on Nvidia and Palantir (Fortune, Sherwood, Yahoo Finance) and the widely cited ~$1.1 billion notional figure.
- Notional vs. premium, clarified: Burry’s own X posts stating he spent about $9.2 million on ~50,000 Palantir puts, and the Motley Fool analysis “But It’s Not What You Think” on the distinction.
- Winding down: CNBC and Bloomberg reporting on Scion’s SEC deregistration in November 2025 and the launch of Burry’s Cassandra Unchained newsletter.
- The skepticism, formalized: Gecko, Fooled by Randomness — why a single legendary trade proves less than it appears, and why you should distrust the loud number.
Frequently asked questions
An American investor and former physician who founded Scion Capital and later Scion Asset Management. He’s best known for correctly identifying the subprime mortgage bubble and shorting it with credit default swaps before the 2008 crisis — the story of Michael Lewis’s The Big Short. He’s known for deep-value, contrarian, high-conviction positions and a willingness to sit in a losing trade until it pays.
Scion’s Q3 2025 disclosure showed put options on Palantir (5 million shares) and Nvidia (1 million shares), reported as a ~$1.1 billion bet against AI. But that’s the notional value, not the money at risk: Burry said he spent about $9.2 million on roughly 50,000 Palantir puts, and warned that “sometimes, we see bubbles… sometimes, there is something to do about it.”
A put reported at notional value can look like a billion-dollar wager when the premium actually spent is a small fraction of that. For a put buyer, the maximum loss is the premium, not the notional — so Burry’s ~$9.2 million is the real risk, while $1.1 billion just describes exposure. Confusing the two is one of the most common errors in reading famous investors’ filings.
He was spectacularly right on subprime, but his record since is mixed and he is frequently early — which in real time is indistinguishable from wrong. He’s made bearish calls that didn’t pan out and closed positions that later moved his way. Being early is the contrarian’s central hazard, which is why his use of defined-risk options — where the loss is capped — matters more than any single call’s direction.
Trader profile in Gecko’s trading psychology series. Details of Burry’s subprime trade are drawn from Michael Lewis’s The Big Short and public record; his 2025 positions, the notional-versus-premium figures, Scion’s deregistration, and the Cassandra Unchained newsletter are drawn from SEC filings and contemporaneous reporting (Fortune, CNBC, Bloomberg, Sherwood, Motley Fool) as of July 2026. Figures are approximate. This profile is not an endorsement of any Burry view or position and takes no stance on the AI trade. Gecko is an educational and informational tool. Nothing here is financial, investment, or trading advice, or a recommendation for or against any security or strategy. Trading carries substantial risk of loss.
Drop in a single statement. Gecko produces a one-page Behavioral Diagnosis ranking your costliest habits in actual dollars. Free to start. No card. No broker connection.
Short notes, usually once or twice a month. Unsubscribe in one click.
No account needed. We use your email only to send Gecko blog posts, and the link at the bottom of every email opts you out in one click.
More on the blog
- 1929 by Andrew Ross Sorkin: The Best Markets Book of 2025 Is a Warning About This OneSorkin's 1929 was the markets book of the year — a #1 NYT bestseller and a Best Book of 2025 across the Washington Post, TIME, The Economist, Bloomberg and more. Built on archival material not seen before, including the New York Fed's board minutes and private diaries, it reconstructs the crash through 75+ figures. Its enduring value for a trader isn't the history. It's the anatomy of the behavior: margin, euphoria, and the certainty that 'this time is different.' The book landed in a market arguing about an AI bubble — and its real lesson is that the mechanics change while the human wiring doesn't.
- GCR: The Anonymous Contrarian Who Shorted LUNA — and What Crypto's Most Famous Trader Teaches About DisciplineGCR (GiganticRebirth) is crypto's most famous contrarian. In March 2022 he escrowed $10M in a public bet against Terra's Do Kwon that LUNA would trade lower in a year; he'd also shorted LUNA and reportedly covered near $0.72, weeks before the collapse to zero. The copyable lessons are behavioral — patience, selectivity, fading sentiment, de-risking winners — not the mythic returns. Read the legend with skepticism: he's anonymous, the '$1K to $1B' figures are unverifiable, and his reported 120-hour weeks are a warning, not a model.
- Trading the Headline Presidency: What Policy-by-Announcement Does to VolatilityThe April 2025 'Liberation Day' tariffs triggered the largest two-day loss in market history (~$6.6T), sent the VIX above 50, and were followed by one of only four sub-100-day drops of VIX from 50 to under 20 on record. The market recovered fully and finished 2025 higher. For a trader, the lesson isn't a market call — it's a regime: policy-by-announcement means headline risk, and the panic-sell-then-FOMO-buy round trip is the costliest response to it. Party-neutral: market mechanics, not politics.