'I can't short OpenAI': Burry reveals bet against another AI company

Investor Michael Burry, who holds a skeptical view of the rapid development of AI, said in a question-and-answer session on his blog that he bet on Oracle's stock falling. He also explained why he shorted Nvidia but did not bet against Meta and other tech giants. Oninvest provides a translation of his answer in its entirety.

Q: You say Nvidia doesn't have a depreciation problem, but hyperscaler companies do, and therefore this could affect Nvidia (Burry argued that bigtechs that buy Nvidia hardware artificially inflate their financials because they underestimate the amortization on AI chips in their reporting. - Oninvest). If that's the case, why would you choose to short Nvidia rather than a hyperscaler like Oracle?

Burry's answer: If I'm betting against Meta, I'm also betting against its dominance of social media and advertising. If I'm betting against Alphabet, I'm betting against Google Search in all its forms, Android, Waymo and so on. If I'm betting against Microsoft, I'm betting against the world leader of office SaaS productivity solutions.

The largest companies are not pure bets against AI. These three giants aren't going anywhere. They will realize they shouldn't have spent so much money, write down assets, and probably revise earnings numbers. But they will still remain globally dominant companies outside of AI development.

I can't short OpenAI, but I would do so at its $500 billion valuation (OpenAI reached this valuation in the fall and in December was already discussing a new round with a $830 billion valuation - Oninvest).

I have put options on Oracle (such an option usually generates income if the price of the underlying asset falls. - Oninvest). I have also been shorting its stock directly for the last six months. I don't like the way this company is positioned and the investments it is making. It didn't need to do what it's doing and I don't understand why it's doing it. Perhaps it's ego.

However, Nvidia is just the cleanest bet [against AI], and perhaps along with its "protégé" CoreWeave. Nvidia is now completely dependent on hyperscaler spending, and I don't see how that math can add up. Nvidia will sell $400 billion worth of chips this year, and there will be less than $100 billion worth of application-level use cases.

I refer you back to Cisco again - and now also to power. I've seen this before and read about all the other cases where this has happened.

Also, Nvidia is the most favored and least doubted company. Therefore, it is cheap to short it, and its put options are cheaper than some of the other major shorts, about which there is much more doubt.

***

Burry became famous after he successfully predicted the collapse of the mortgage market in the United States, which led to the global financial crisis of 2008. His story was described in the book "The Downgrade Game", which was the basis for the movie of the same name. In 2021 Burry warned that the market was forming "the greatest speculative bubble of all time in all areas".

In the third quarter of 2025, Burry's hedge fund Scion Capital purchased approximately 80% of its portfolio in put options on shares of chipmaker Nvidia and AI developer Palantir. Burry closed Scion Capital shortly thereafter, shrugging off the obligation to disclose his trades. In December, Burry also called Tesla shares "ridiculously overvalued" but said he was not shorting them.

This article was AI-translated and verified by a human editor