According to Business Insider, Michael Burry’s Scion Asset Management purchased bearish put options on 1 million Nvidia shares and 5 million Palantir shares last quarter, with notional values of $187 million and $912 million respectively. These bets dominated the firm’s US stock portfolio, which contained only eight total holdings worth just $68 million in direct positions. Burry returned to X after a two-year hiatus to warn about AI hype, updating his profile to “Cassandra Unchained: Missteps to Mayhem, Coming December 2025.” He posted charts showing slowing growth in cloud computing divisions at Amazon and Alphabet, surging tech capital expenditures reminiscent of previous bubbles, and circular dealmaking between AI companies. Following the disclosure, Nvidia fell 4% and Palantir dropped 8%, with Palantir’s CEO Alex Karp publicly questioning why Burry would short companies that are “making all the money.”
Bubble warnings everywhere
Burry’s not exactly being subtle here. He’s basically screaming “bubble” through movie references and historical parallels. The “WarGames” quote about the only winning move being not to play? That’s about as direct as you can get without actually saying “sell everything now.” And his Cassandra reference is telling – he clearly feels like he’s shouting into the void while nobody listens.
But here’s the thing: his charts actually make some sense. Cloud growth slowing while everyone’s dumping billions into AI infrastructure? That’s exactly what happened during the dot-com bubble when companies built massive fiber networks that sat dark for years. And the circular dealmaking between Nvidia, Microsoft, OpenAI and others? It reminds me of that old joke about economists assuming a can opener – everyone’s buying each other’s services and calling it revenue.
Massive bets, massive risks
Let’s be real – betting against Nvidia right now feels like trying to stop a freight train with your bare hands. The company just hit a $5 trillion market cap and seems unstoppable. Palantir at nearly $500 billion? That’s more than companies that actually make consistent profits like Mastercard and Exxon.
But Burry’s playing a dangerous game here. As Russ Mould pointed out, the “danger” is that Burry “gets run over by momentum and liquidity-fueled markets” if these stocks keep rising. And with the Fed potentially cutting rates? That could pour gasoline on this already raging fire. These aren’t small positions either – we’re talking over $1 billion in bearish bets concentrated in just two stocks. That’s either incredibly brave or incredibly foolish.
Historical echoes are loud
What’s fascinating is how similar this feels to 1999. Daniel Bustamante nailed it when he said we have “all tinder soaked in gas” with retail investors crowded into the same names and margin debt at all-time highs. Remember when Cisco was the Nvidia of its day? Everyone thought it couldn’t fail until it lost 80% of its value.
The telecom boom and bust that Burry referenced in his book excerpt is particularly chilling. Companies built massive infrastructure that nobody needed, prices collapsed, and high-flyers went bankrupt. Sound familiar? We’re seeing the same pattern with AI data centers and chip factories being announced weekly. What happens if demand doesn’t materialize as expected?
Is not playing the winning move?
So where does this leave ordinary investors? Burry’s essentially saying the smartest move might be to sit this one out entirely. But that’s easier said than done when everyone around you is getting rich (on paper, at least).
The real question is timing. Burry was early on the housing crash – famously so. He started betting against subprime in 2005, and had to endure two years of pain before being proven right. Could we see the same pattern here? AI might be overhyped, but that doesn’t mean it’s going to collapse tomorrow. These things can stay irrational longer than you can stay solvent, as the old saying goes.
One thing’s for sure: when Michael Burry makes moves this bold, it’s worth paying attention. Even if he’s wrong, his reasoning exposes real vulnerabilities in the current AI narrative. The next year should be absolutely fascinating to watch.
