Michael Burry is back on the other side of the trade. Scion Asset Management disclosed sizable put positions tied to Nvidia and Palantir, days after the hedge fund manager warned that AI spending now resembles the excesses of the dot-com era. Regulatory filings show Scion holds put options referencing roughly 5 million Palantir shares and about 1 million Nvidia shares, with notional exposure of about 912 million dollars and 186 million dollars, respectively. The timing hit a fragile tape. Palantir slid more than 7% in premarket trading despite raising its full-year sales outlook after a strong quarter, while Nvidia dipped nearly 2% even after a year-to-date rally of 54%. The message from Burry is clear: the AI trade’s momentum is colliding with valuation math, and the market suddenly cares.
Scion’s latest quarterly filing puts two of the highest-profile AI winners squarely in its sights. The fund’s bearish wagers against Palantir Technologies and Nvidia Corp. appear to express a view that AI leaders have run too far, too fast. The disclosure follows Burry’s recent warning that AI-related capital expenditures now rival the intensity of the 1999–2000 tech bubble, a comparison that has resonated across trading desks. While filings do not reveal strike prices, maturities, or whether puts are directional bets or hedges against other positions, the notional share counts are large enough to matter. They also skew heavier toward Palantir, underscoring where Burry seems to see the most excess. Against a backdrop of euphoric positioning in select AI names this year, Scion’s move reads as a challenge to the market’s favorite narrative.
The immediate reaction was a pullback in the most crowded corners of AI. Palantir fell sharply in early trading, while Nvidia’s premarket dip added to pressure on broader equity futures. The setup is notable: both companies are coming off strong prints and substantial year-to-date gains, but the incremental buyer is getting scarce. Momentum strategies that chased the past leg higher now contend with headline risk and a rising chorus of caution from heavyweight Wall Street executives. That’s made the first signs of distribution in AI leaders more consequential for indices, factor baskets, and retail-favorite lists alike. The question isn’t whether AI is real. It’s whether the next dollar of revenue and profit growth will justify another leg up in already demanding valuations. Burry’s disclosure pushes that question to the front of the order book.
Palantir gave bulls a lot to work with. Third-quarter revenue rose 63% to about 1.2 billion dollars, net income climbed 40% to 476 million dollars, and management lifted its full-year revenue guide to a range near 4.4 billion dollars. By any normal standard, those are impressive numbers—especially given the mix of U.S. government work and accelerating commercial AI demand. Yet the stock sold off, a sign that the market was positioned for more or had grown uneasy with the price tag. Nvidia, meanwhile, remains the poster child of AI scale. The company has turned hyperscaler capex into a profit engine and delivered one strong quarter after another. Even so, a modest premarket decline after Scion’s filing reinforces how tightly wound expectations have become. When good is no longer good enough, downside gap risk grows.
Burry’s skepticism is landing as Wall Street leaders warn of a broader market pullback. Top bank CEOs have floated the risk of a 10% or larger correction over the next two years, citing stretched valuations and the uncertain payoff timeline for massive AI investments. That dovetails with Burry’s view that AI capex has begun to mirror dot-com-era spending—a boom that produced real technology but punished equity holders who paid peak multiples. The hyperscalers can afford the bet, but the earnings conversion for the broader economy is still being proved. If the market starts treating AI spend as a cyclical capex story rather than a straight-line profits story, multiples at the front of the trade will compress. In that environment, even flawless execution by Nvidia and Palantir might not shield their share prices.
It’s worth underscoring how to read this filing. The reported share counts in 13Fs reflect notional exposure, not option premium paid, and the dollar values often map to the underlying, not the cash at risk. Strike details, expiration dates, and whether puts hedge other longs are not disclosed. In other words, the headline notional is big by design, but the check Scion wrote could be far smaller. That caveat has tripped up investors before—Burry has used puts both as expressions of macro caution and as portfolio insurance. Still, scale matters. Tying up options on 5 million Palantir shares and 1 million Nvidia shares is a deliberate signal. The asymmetry of those two figures also hints at where Scion perceives the greatest gap between story and valuation. If momentum fades, liquidity will decide how quickly that gap closes.
For Nvidia, the bull case is about insatiable demand for AI training and inference, a supply chain it dominates, and product cycles that keep competitors chasing. For Palantir, it is about converting a first-mover edge in applied AI into sticky government and commercial contracts at rising margins. Neither story is broken. The bar, however, has risen. To extend gains from here, Nvidia needs the hyperscalers and enterprise buyers to reaffirm multi-year capex plans with durability through macro noise. Palantir needs to keep compounding large deal wins without margin slippage and show that AI platforms are not a one-quarter wonder. Any wobble—on bookings, pricing, or deployment timelines—will validate the skeptics and strengthen the hand of trades like Scion’s. The market is shifting from believing to verifying.
Catalysts ahead will stress-test both sides. The next wave of earnings from cloud providers will clarify whether AI spending is accelerating, plateauing, or getting reallocated. Guidance from chip suppliers and system integrators will color the downstream demand picture for Nvidia. Contract disclosures and commercial customer additions will steer Palantir’s visibility. Macro also looms: a hotter inflation print or a renewed rise in long-term yields would compress multiples across tech, while a benign inflation path and a steadier Fed could cushion any AI-related drawdowns. If AI deployments begin to show measurable returns in core business metrics—cost saves, new revenue, clear productivity gains—valuation concerns will ease. If not, the market will revert to paying less for growth that looks more cyclical than secular. Burry’s puts will look prescient in that case.
Strip away the drama and the trade says this: positioning and price are fragile when narrative outruns fundamentals. Burry is challenging the consensus that AI leaders are untouchable, and the first-day reaction suggests that challenge is landing. Even strong quarters could be met with sell-the-news behavior if investors conclude the future is already in the price. That puts Nvidia and Palantir at a pivotal moment, not because their businesses cracked, but because the market’s willingness to keep extrapolating is in question. If the AI capex wave keeps compounding and turns into broad-based profit growth, this pullback will look like noise. If spending stalls or payoffs lag, the downside can come fast. For now, the AI trade is back in price discovery—and one of its most famous skeptics just forced the issue.
