The Big Short was probably the most popular investment-related movie of all time (right up there with Wall Street in the 1980s), and it put a spotlight on Michael Burry, a hedge fund manager who pretty much discovered and nailed the subprime bubble in 2007/2008, bet big against it and made a killing—and, mostly thanks to the movie, a following.
Admittedly, I don’t really keep tabs on more than a few money managers out there (a couple of which I touch base with here and there), but every now and then Burry’s moves grab the headlines—mostly (not always) on the bearish side, often looking for good-sized drops in certain stocks, Bitcoin or the market as a whole; like everyone, some of those moves have been decent (he was worried and short in 2021, a bit early but it did lead to the big 2022 growth bear) and some haven’t (he was short in the summer of 2023, early in a fresh bull market).
I’m mentioning all of this today because, this week, it was revealed that Burry has a $1 billion-plus bet against two of the biggest beneficiaries of the AI wave—Nvidia (NVDA) and Palantir (PLTR)—likely as a bet against AI as a whole. Will he be right? Let’s put both stocks under a microscope and see what we find.
[text_ad]
Nvidia (NVDA)
We’ll start with Nvidia, which has been the flag-bearer of the AI wave since early 2023, when it initially took off on a jaw-dropping earnings report that spring—suffice it to say that the firm’s AI chips (and, increasingly, bundled networking offerings) are the core of the AI buildout.
And while growth has slowed some from the prior triple-digit pace, it remains outstanding—sales in the quarter ended July boomed 56% while earnings were up 54%, and analysts don’t see much deceleration from here, with the top and bottom lines expected to rise 50% (give or take) for many quarters to come.
Moreover, that’s likely to prove conservative—at a recent industry conference, the company released a few tidbits about cumulative orders for its Blackwell and Rubin AI chips that have many analysts thinking even next year’s buoyant targets could prove very conservative. Even going with today’s views, the stock trades at “only” 30x 2026 estimates—not unreasonable given the growth.
But what about the stock? I actually think it looks very good here, partly because NVDA etched what we dub a “re-set” correction—shares essentially started to top out in June 2024 (they made new peaks in the fall by a bit, but relative to the market, they didn’t) and then went over the falls with everything else earlier this year; at its lows, NVDA had made no net progress for over a year. To me, that probably re-set the longer-term advance, wearing out the weak hands over time.
Confirming that thesis was the rebound from the lows, which was unusually persistent (up 11 weeks in a row and 15 of 16), telling you big investors were consistently adding. More recently, too, the stock popped out of a two-plus-month rest period, too. I actually think NVDA looks good around here—a drop below 175 would have my antennae up, but a move back above 212 or so would be confirmation the uptrend is in solid shape.
Palantir (PLTR)
Then there’s Palantir, which has the hands-down leading software platform for AI inference, which allows firms to actually get major value from those trained AI models—letting them access a firm’s data in a safe, secure way, and then providing actionable information (or just taking an action itself) to boost a client’s efficiency.
Beyond the story, the numbers here are fantastic—sales (up 63% in Q3) and earnings growth (up 110%) continue to accelerate, with U.S. business (especially commercial clients, up 121%!) leading the way. And with a backlog of $3.63 billion (up 199%!) from those U.S. commercial clients, growth will obviously remain rapid for at least the next few quarters.
All that said, the chart does have some yellow flags, the first simply being that PLTR has had a monstrous move without any prolonged rest period—unlike Nvidia. From PLTR’s breakout near 30 last summer, shares are up eightfold, which makes the stock far more vulnerable to some bad news. Indeed, shares trade at around 200x 2026 earnings estimates; valuation isn’t a huge factor for me, but it’s worth noting we’re clearly in nosebleed territory.
We’d also note that the stock really hasn’t made any net progress since early August and buying has been tepid—the last week of above-average-volume buying was at the early-August peak, and the one prior to that came at the market bottom.
Are those reasons to think Michael Burry will be right? Not necessarily, but here’s what I would say: A decisive drop below 170 (the stock’s low area in recent weeks) would make the past three months look more like a top and could usher in a more prolonged correction/consolidation. So far, it’s OK, but worth watching.
So will Michael Burry’s bearish bets be proven right or wrong?
If you’ve been following me for any length of time, you know I’m not big on predictions—I much prefer to be flexible and take it as it comes. Still, if we’re playing the guessing game, I’ll simply say that PLTR looks much higher risk to me than NVDA: a later-stage advance, a negative reaction to earnings this week and a valuation that makes even the growthiest growth investor blush.
What are your thoughts? Are PLTR and NVDA set to decline along with the AI sector? Or is Burry barking up the wrong tree again? I’d love to hear your thoughts at mike@cabotwealth.com.
[author_ad]
