Michael Burry, the investor who became a household name for betting against the US housing market before the 2008 financial crisis, has sold his December 2026 put options on Nvidia entirely. The move is part of a broader portfolio trim that also saw him exit his near-term Palantir puts, though his overall bearish posture on AI stocks remains very much alive.
The September 9 adjustment looks less like a change of heart and more like a chess player pulling a piece back to reposition. Burry still holds long-dated December 2027 puts on Palantir and the Nasdaq-100 (QQQ), and his short stock positions account for more than 21% of his portfolio even without counting put options.
What Burry actually did, and what he didn’t
Burry chose not to roll these near-term options into later-dated contracts, instead opting to free up capital. He’s described the move as a way of moderating exposure while keeping liquidity available to capitalize on what he anticipates will be an “interesting market this fall.”
Nvidia remains a prominent short position in his portfolio, sitting alongside Oracle, Palantir, and Nebius. His largest long positions tell their own story about where he sees value. Lululemon accounts for roughly 17% of his holdings, followed by MercadoLibre at approximately 12% and Molina Healthcare rounding out the top three.
The AI skeptic’s thesis
His critique centers on two pillars. First, the sustainability of AI capital expenditures. Companies are pouring billions into GPU purchases and data center buildouts, and Burry questions whether the revenue these investments generate will ever justify the spending.
Second, he’s taken aim at the financing structures that enable these massive GPU purchases. Burry has described some of these arrangements as “fugazi,” a word that roughly translates to “fictitious” or “fake.” His concern is that the financial engineering propping up AI infrastructure spending may be masking the true economic picture.
He’s also raised the specter of semiconductor obsolescence, arguing that the breakneck pace of chip development means today’s cutting-edge hardware could become tomorrow’s paperweight faster than investors expect.
What this signals for AI investors
By trimming near-term puts and maintaining longer-dated bearish bets, he’s essentially saying: the reckoning might not come this quarter, but I still think it’s coming.
The short positions exceeding 21% of his portfolio serve as a data point worth considering. The concentration of his long book in consumer retail and healthcare also carries an implicit message about relative value, as Burry is finding more attractive risk-reward in sectors that don’t require faith in exponential technology adoption curves.
His comment about anticipating an “interesting market this fall” is deliberately vague, but the liquidity positioning speaks louder than words. Freeing up capital ahead of expected volatility is a move that suggests he wants dry powder ready when opportunities emerge.