So here's something that's been catching my attention this week: the guy who called the 2008 subprime crisis—the actual person behind 'The Big Short'—is basically doing the opposite of what everyone else in the market is doing right now. And he's doubling down on it.

Michael Burry just disclosed his latest moves, and they're wild. While Nvidia's market cap is sitting at $5.3 trillion and the Nasdaq keeps hitting new highs, he's actively shorting semiconductor ETFs, Nasdaq ETFs, and maintaining massive put positions on both Nvidia and Palantir. But here's the interesting part: he's simultaneously buying up traditional software companies that have gotten absolutely hammered by the AI narrative—Adobe, Autodesk, Salesforce, Veeva Systems. It's a complete flip-the-script portfolio.

Let me break down what he's actually betting on. Back in Q3 2025, Burry loaded up on puts—about $912 million in notional value on Palantir and $187 million on Nvidia (though the actual cash deployed was closer to $9.2 million, which is a detail Wall Street got hilariously wrong). That move shocked the market. But he didn't stop there. He closed his hedge fund, went independent, and now he's publishing his thesis on Substack under the column 'Cassandra Unchained'—which, if you know your mythology, is basically him saying nobody's going to believe him anyway.

The Palantir short is actually working out. The stock's down from $161 when he entered to around $137 now, even after they just posted an 85% YoY revenue beat. Burry didn't take profits though. He's holding puts with strike prices at $100 (expiring December 2026) and $50 (expiring June 2027), which means he thinks the stock could drop another 60% from here. He's literally saying Palantir's fair value is 'low single to double digits.' In April, he posted that Anthropic was 'eating Palantir's lunch' with better AI tools, and the stock dropped 13.7% that week. He deleted the post later, but the damage was done.

Nvidia's a different story though. The stock's near all-time highs at $215, and his puts are underwater. But instead of cutting losses, he's actually increased his position. His core thesis is 'over-investment in AI infrastructure'—basically saying this mirrors the dot-com bubble. He compares Nvidia to Cisco, which went up 3,800% before the crash, then fell 80%. His argument: companies like Microsoft, Google, Meta are extending GPU depreciation periods to juice their financials, and between 2026-2028 this will understate depreciation by roughly $176 billion across the industry. Nvidia fired back with a seven-page memo to Wall Street analysts directly addressing his claims, which is honestly kind of telling.

But the most fascinating part of his portfolio isn't the shorts—it's the longs. Adobe, Autodesk, Salesforce, Veeva, MSCI. These are solid companies that have gotten absolutely wrecked because the market decided they're 'AI losers.' Adobe's down 30% from its 52-week high, Autodesk is down 22% this year. Their valuations have crashed back to 2018-2019 levels. Burry's thesis: the market has massively overcorrected. These aren't broken companies; they're just out of favor.

So what he's really constructed here is a hedge: if the AI bubble narrative cracks, the highly valued AI beneficiaries (Nvidia, Palantir) crater first, while the beaten-down software stocks recover. It's the inverse of 'The Big Short'—same contrarian playbook, different direction.

Here's the thing that gets me: Burry himself said 'My judgment of securities value has been out of sync with the market for a long time.' That's not a complaint—that's literally his brand. He's built a career on being early and being alone. Whether this particular bet works out or not, the meta-narrative is worth watching. The crowd is all-in on AI, and he's standing on the complete opposite side of the room.