Michael Burry, the investor who famously predicted the 2008 housing crash, disclosed in his final SEC filing that he'd placed roughly $1.1 billion in notional bearish bets against Nvidia and Palantir, and by mid-2026 had reportedly expanded that position to as much as 80% of his portfolio while adding new short positions against Oracle and the semiconductor sector broadly. His public reasoning connects directly to the accounting concerns and circular-financing structure covered earlier in this series. This is a factual account of his disclosed positions and stated reasoning, not investment advice or a prediction of who's right.
The Original Bet: Palantir and Nvidia Puts
Burry's final verified regulatory filing, a 13F covering the quarter ending September 30, 2025, showed his hedge fund Scion Asset Management holding put options with a notional value of about $912 million against Palantir and $187 million against Nvidia, alongside a handful of other positions including calls on Pfizer and Halliburton. The disclosure prompted an unusually pointed public reaction from Palantir CEO Alex Karp, who called Burry's wagers "super weird" and used considerably harsher language in describing them. By June 10, 2026, at least one leg of the bet looked directionally correct: Palantir shares, which traded at $200.47 at the end of Q3 2025, had fallen to $132.07 by early June, a decline of roughly 34%.
Doubling Down in 2026
Burry deregistered Scion Asset Management with the SEC in November 2025, announcing plans to liquidate the fund and return capital to investors — which means everything he's disclosed since then comes through his own voluntary public posts rather than an audited regulatory filing, a distinction worth keeping in mind. In a May 2026 post, he disclosed that as much as 80% of his portfolio was allocated to Nvidia and Palantir put options, with an estimated notional value near $1.1 billion — roughly double the size of his last officially filed position. The specifics he shared: put options tied to 1 million shares of Nvidia with a $110 strike price expiring in 2027, and a layered short position on Palantir combining puts at $100 (expiring December 2026) and $50 (expiring June 2027) strike prices. He also disclosed expanded bearish bets against the Philadelphia Semiconductor Index, the Nasdaq 100, and Oracle specifically — the same company whose stock wobbled on circular-financing fears covered earlier in this series — together representing roughly 9.5% of his total short exposure.
Why He's Betting Against It
Burry's stated reasoning connects directly to themes already covered in this series. In November 2025, he publicly accused major AI hyperscalers of artificially boosting their earnings — an accusation that echoes the accounting questions raised by Microsoft's data-center depreciation schedule change and the equity-stake gains inflating Amazon's and Alphabet's reported profits, both covered in earlier entries here. On Substack, he described a more qualitative concern too: after a long drive spent listening to financial broadcasts, he wrote that he'd noticed "no end to the discussion of AI," describing a self-reinforcing market narrative that was increasingly displacing fundamental financial analysis in how investors were pricing these companies.
The Fine Print Headlines Skip
The eye-catching "$1 billion bet" framing that accompanies most coverage of Burry's position overstates the actual capital at risk in a way worth understanding. An options analysis of his Palantir position specifically found that buying puts on a stock trading around $140 with a $50 strike price requires the stock to fall by roughly two-thirds before those specific contracts pay off at all — meaning the real premium Burry spent to place that particular bet is a small fraction of the headline notional value being widely quoted. It's also worth noting Burry isn't broadly betting against the market or technology as a category: alongside his AI-specific shorts, he's disclosed long positions in Microsoft, MSCI, PayPal, Adobe, Mercado Libre, and Lululemon — a targeted bet against a specific set of companies and valuations, not a blanket rejection of tech stocks generally.
He's Not Alone
Burry isn't the only prominent investor expressing skepticism. Peter Thiel fully exited his Nvidia position, and SoftBank sold its entire $5.8 billion Nvidia stake specifically to help fund its investment in OpenAI — a move that, notably, feeds directly into the circular-financing web covered in the previous entry of this series, where the money changing hands between these companies increasingly flows in a closed loop rather than from clearly independent sources.
The Honest Counter-Case
Burry's track record with the 2008 housing crash lends his current position real credibility, but it doesn't guarantee he's correct this time, and the bull case remains substantial. Hyperscaler capital spending has continued accelerating rather than pulling back, with the combined 2026 figures covered earlier in this series reaching roughly $725 billion across the four largest cloud providers, and Microsoft specifically signed new multibillion-dollar infrastructure deals with Nebius and Iren during the same period Burry was expanding his short positions. Betting against a trend this well-capitalized has been a losing trade for plenty of skilled investors before, and the outcome of Burry's specific options positions, most with 2026 and 2027 expiration dates, remains genuinely undetermined as of this writing.
Burry's Disclosed AI Positions at a Glance
| Position | Reported notional value | Structure | Source |
|---|---|---|---|
| Palantir puts (original) | ~$912 million | Put options, strike/expiration undisclosed in filing | 13F filing, Q3 2025 (SEC-verified) |
| Nvidia puts (original) | ~$187 million | Put options, strike/expiration undisclosed in filing | 13F filing, Q3 2025 (SEC-verified) |
| Nvidia puts (expanded) | Tied to 1M shares | $110 strike, expiring 2027 | Voluntary disclosure, May 2026 (not SEC-audited) |
| Palantir puts (expanded) | Part of ~$1.1B total notional | $100 strike (Dec 2026), $50 strike (June 2027) | Voluntary disclosure, May 2026 (not SEC-audited) |
| SOXX, QQQ, Oracle shorts | ~9.5% of total short exposure | Not fully disclosed | Voluntary disclosure, May 2026 (not SEC-audited) |
Frequently Asked Questions
Is Michael Burry's $1 billion AI short bet actually accurate?
The $1 billion figure refers to the notional value of his put options, not the actual capital he spent. Options analysis suggests the real premium paid for positions like his Palantir puts is a small fraction of that headline number, since the stock would need to fall significantly before the contracts pay out.
Does Michael Burry still file official SEC reports on his positions?
No. He deregistered Scion Asset Management with the SEC in November 2025 and announced plans to liquidate the fund. Everything disclosed since then comes through his own voluntary public posts, which are not independently auditable the way a 13F filing is.
Why does Burry think AI stocks are in a bubble?
He has publicly accused major AI hyperscalers of artificially boosting their earnings and described a self-reinforcing market narrative around AI that he believes is displacing fundamental financial analysis in how investors price these companies.
Is this article investment advice?
No. This is a factual account of Michael Burry's publicly disclosed positions and stated reasoning, not investment advice. Anyone considering an investment decision should consult primary financial filings and a licensed financial advisor.