In a decisive move that underscores his reputation as a staunch contrarian, Michael Burry’s Scion Asset Management has taken a massive bearish stance against the market’s leading artificial intelligence companies. The firm’s most recent 13F filing reveals that put options on Palantir Technologies and NVIDIA constitute nearly 80% of its reported portfolio value, representing a high-conviction bet that the AI trade is overextended.
This positioning signals a firm belief that current valuations for AI-centric stocks are unsustainable. Below is a detailed analysis of Scion’s top holdings for the quarter.
The portfolio’s core consists of two dominant bearish positions:
Palantir Technologies Inc. constitutes 66.04% of the portfolio through a put option on 5,000,000 shares with a notional value of $912.1 million. This position indicates deep skepticism toward Palantir’s valuation and the AI narrative surrounding the company.
NVIDIA Corp. represents 13.51% through a put option on 1,000,000 shares valued at $186.6 million. This bet against the chipmaker at the heart of the AI infrastructure boom reflects Burry’s historical skepticism of market exuberance.
Alongside these substantial bearish wagers, Scion maintains smaller, tactical bullish positions in undervalued sectors:
Pfizer Inc. call options account for 11.07% of the portfolio, valued at $152.9 million, suggesting a contrarian play on the pharmaceutical company’s depressed valuation and dividend yield.
Halliburton Co. call options represent 4.45% at $61.5 million, providing leveraged exposure to potential rebounds in energy prices and oilfield activity.
The fund also holds straight equity positions in Molina Healthcare and Lululemon Athletica, alongside new positions in SLM Corp. and Bruker Corp. preferred convertible shares.
Scion completely exited several major holdings, including UnitedHealth Group, Regeneron, Meta Platforms, Estée Lauder, JD.com, and Alibaba, indicating a strategic pivot away from prior investment themes.
This portfolio construction reveals a sophisticated strategy: substantial bearish bets against overhyped AI stocks are complemented by tactical bullish positions in undervalued cyclical and defensive sectors. The approach demonstrates Burry’s continued focus on asymmetric risk-reward opportunities and his willingness to take significant contrarian positions against prevailing market narratives.
For all the latest news and podcasts, join our free newsletter here.
Don’t forget to check out our FREE Large Cap 1000 – Stock Screener, here at The Acquirer’s Multiple:
4 Comments on “Michael Burry’s Scion Asset Management Q3 2025 Portfolio Analysis”
What kind of article is this, is it written by some dumb AI tool? He has bought put options in Palantir and Nvidia which means he thinks these companies are overvalued in public market, not a conviction in those companies. Also, widely reported everywhere, the value of those holdings are calculated based on underlying shares, not necessarily premium he paid to but those put options, hence the actual value of put options would be very miniscule in his portfolio. It is very low article that doesn’t belong to your website.
Thanks for you comment — the article does list both Palantir and NVIDIA as (Put) positions, exactly as shown in the filing. The analysis focused on the portfolio composition and size of exposure rather than Burry’s directional intent. You’re right that 13F values reflect the notional exposure, not the option premium — appreciate you highlighting that distinction.
Johnny. is anyone else bewildered at Burry rolling in depreciation on computing hardware to 2 years? When GAAP and external auditors are all fine with tech companies using 5 year schedule for computing write downs. Meta and similar companies is in complete compliance with accounting rules. But burry moved the goal posts to frame his argument. What tech company would go to the trouble of installing thousands of GPU’s into data centers only to rip them all out 2 years later. Even if you took Burry on his word and accepted his logic, then why would you open a short on the pick and shovel play Nvidia. Burry makes no sense this time. (And I am a big fan of his). GLTA
You’re right that no company plans to rip out hardware after two years. Burry’s short-term depreciation argument isn’t about accounting rules, but about economic reality. He’s betting the AI chip upgrade cycle will be brutally fast, making current GPUs obsolete long before their 5-year accounting life.
This connects directly to his NVIDIA short. The “pick and shovel” analogy fails if the gold rush is a bubble. Burry is betting we’re at peak AI capital spending. If companies don’t see the expected returns, their spending will collapse, and NVIDIA’s sales will fall off a cliff—even if they are the “picks and shovels” seller.
His trade is a bet that the AI investment boom is unsustainable, hurting both the users (via rapid obsolescence) and the supplier (NVIDIA) simultaneously.