Michael Burry's puts on AI stocks: what the positions show

What did the filing on Michael Burry’s puts on AI stocks report?

A regulatory filing for the quarter ending 30 September 2025 listed put options held by Burry’s fund, Scion. The notional value was about $187 million on Nvidia and $912 million on Palantir (CNBC, November 2025). Those two lines are the entire public record of the positions, and everything else in the press coverage is interpretation.

The filing did not state strike prices or expiry dates. It gave a snapshot at one date, and it did not say whether the positions were later sold, rolled, or left to expire. A careful reader keeps the snapshot and the story apart, and writes down which sentence comes from which.

What is a put option, and how do Michael Burry’s AI puts use it?

A put option gives its owner the right to sell a stock at a set price before a set date. The owner gains if the price falls below that level and loses the premium if it does not. Puts are the standard tool for expressing a falling-price view with a capped loss, which is why a bubble sceptic would use them.

The cap is the useful feature. Selling a stock short can lose more than the amount invested, while a put loses at most the premium. That makes puts a measured way to be early, and the price of being early is a premium that decays each day until the option expires.

What does notional value mean in Michael Burry’s AI puts?

Notional value counts the stock that the options cover, at the current share price. It is the size of the exposure, not the money at risk. As an illustration with invented numbers, a put covering $1,000,000 of stock might cost a premium of $50,000. The loss is capped at $50,000, while the notional figure is 20 times larger.

So $912 million of notional on Palantir does not mean Burry risked $912 million. The real cash at risk is a fraction, and the filing does not give it. Headlines that quote notional as if it were a wager overstate the risk taken and the size of the conviction.

Why did Michael Burry’s puts name Palantir and Nvidia?

The filing names two companies of different kinds. Nvidia sells the chips at the centre of AI spending. Palantir sells software and analytics, and its share price carries high expectations of future growth. The Palantir notional of $912 million is almost 5 times the Nvidia figure of $187 million, so the larger line sits on the software name.

Burry has not been quoted in the sources checked for this page as explaining the choice, so reasons stay speculation. What the data does support is that the exposure was not limited to the chip maker. The positions spread across hardware and software, which fits a view about AI valuations in general.

Do Michael Burry’s puts on AI stocks prove a bearish bet?

A put can be a bet or a hedge. A fund that owns other AI shares might buy puts to cap its losses, and the filing would look the same. CNBC noted that it is unclear whether the positions are still held or served as a hedge. The data cannot settle the question.

A fair reading is that the puts are consistent with scepticism and do not prove a bearish bet by themselves. Proof would need the full portfolio, strike prices, and expiry dates. Until then, the filing counts as one clue that sits beside the depreciation argument, not as a verdict on the AI bubble.

How can you read an option filing like Michael Burry’s in 4 steps?

Step 1: find the date, because a snapshot ages quickly. Step 2: read the notional value and divide by the share price to see the number of shares covered. Step 3: look for strike price and expiry, and note when they are missing. Step 4: ask what else the fund holds, since a put next to a large long position is a hedge.

Apply the four steps to the Scion filing and you reach a clear result: a date of 30 September 2025, two notional values, no strike, no expiry, and no portfolio context in the press coverage. The method leaves you with questions you can state exactly, which beats a conclusion you cannot defend. Reuse the same four steps on the next quarterly filing, and compare the new snapshot with this one to see whether the position grew, shrank, or disappeared.

What should you take from Michael Burry’s puts on AI stocks?

Take the direction, the date, and the limits. A respected investor held puts on two AI names at the end of September 2025. That is a data point about one fund and not a market forecast. It adds weight to the debate about valuations and no weight to any date for a fall.

This page is general information as of October 2026, not investment advice. If an AI exposure inside your business or portfolio needs sizing, the EQOS Asia pricing page lists the diagnostic formats that turn a position into a scored scenario with a method you can recompute.

Frequently asked questions

About the author

Stéphane Jambu

President, EQOS Forecast Asia

Stéphane Jambu leads EQOS Forecast Asia and works as a consultant in e-commerce, search visibility, and AI. LinkedIn profile