7 Big-Name Funds Keep Circling the Same AI Stocks

For AI bulls, Thursday was the kind of day that makes you want to chuck your phone out the window. Lam Research (LRCX) posted its first 40% operating margin in history and still got clobbered by nearly 12%. Meta (META) fell 7% after telling Wall Street it’s spending even more on AI.

Lam and Meta both sit on a list I’ve been building: stocks a group of heavyweight AI funds can’t seem to quit. Every quarter they have to show the SEC what they own on a 13F, and that’s about as close as most of us will get to looking over their shoulders.

So I pulled the last three filings for Brad Gerstner’s Altimeter Capital, Gavin Baker’s Atreides Management, Philippe Laffont’s Coatue, and Alex Sacerdote’s Whale Rock. I also threw in Chase Coleman’s Tiger Global, Stanley Druckenmiller’s Duquesne Family Office, and David Tepper’s Appaloosa.

Druckenmiller and Tepper trade more like macro guys, and they’re two of the best traders on the planet.

The tech guys in this bunch frequently sit across the table from the people developing this technology. Gerstner has interviewed Jensen Huang and Sam Altman on his own podcast and keeps getting called up to fill in on All-In. Baker has been on Invest Like the Best six times.

Going through the filings, I ignored anything under 1% of a fund’s portfolio and focused on stakes above 2.5%. I care a lot less about one fund loving a stock than about four or five of them holding it in size quarter after quarter.

Amazon Is the Easy Part

Amazon (AMZN) was the only stock all seven funds held above 1% in at least two of the three quarters, and it cleared 2.5% in 19 of the 21 filings I looked at. Taiwan Semiconductor (TSM) was nearly as popular, with six funds holding more than 1% in all three quarters. Alphabet (GOOGL), Nvidia (NVDA), and Meta weren’t far behind.

Owning Amazon, Taiwan Semi, or Nvidia isn’t exactly a secret handshake. The good stuff was the next tier down, and it reads a lot like the AI stack series I just wrapped up.

Lam and Applied Materials (AMAT) make the equipment that builds chips, and both kept turning up, along with ASML (ASML) in smaller but steady positions.

Then came the companies that wire it all together, like Coherent (COHR), Ciena (CIEN), and Corning (GLW), which makes the fiber.

Cerebras (CBRS) landed in four of these portfolios, though it only went public in May, so there’s just one filing to go on. The stock now sits well below its $185 IPO price as insider shares unlock. While the company’s technology is reportedly impressive, none of these funds are rushing to buy more.

I’m not sure what level folks are using as fair value, but the stock’s been cut in half over the past month. Once fear around insider unlocks subsides, this stock may start building a base.

Gerstner Was Selling Before the Filing Caught Up

The catch is the timing. A 13F can come out as late as 45 days after the quarter ends, so the June snapshot likely didn’t show up until mid-August. And the guys running these funds are mostly traders at heart, constantly adjusting their positions. You can be sure that some, or even a lot, of what shows up in a filing is already stale by the time we see it.

Look at CoreWeave (CRWV). Altimeter added to the stock in the first quarter. Then, earlier this month, Gerstner said he no longer owned any CoreWeave and had been moving money into Meta and Tesla (TSLA).

We’ve all seen it. CoreWeave shares have taken a beating, and yields on its 2032 notes have hit 15% in the secondary market. Anyone who bought the stock based on an old filing because Gerstner owned it may have bought while he was heading for the exit.

You never want to read a 13F and think, “These guys own it today, I guess I’ll buy it too.”

Treat these filings as a record of what investors who know this industry inside and out kept holding, and how much, from the end of last year through June. When the funds go shopping again, they’ll likely reexamine these stocks first.

AI volatility is insane right now, and the last thing I’d suggest is jumping into any of these stocks Friday morning. But if you believe, as I do, that we’re still in the early innings, this is where I’d start a watch list. Then it’s on us to dig into what each company does, how it stacks up against its competitors, and, most importantly, how much good news is already baked into the price.

Oh, and don’t underestimate the value of a logical technical setup that you can measure your risk against. It’s fine to use fundamentals, 13Fs, and the promise of AI to build a narrative, but I’d rather rely on technicals to tell me when to hit the buy key.

Lam’s Selloff Won’t Show Up Until February

The September filings land in mid-November. They won’t show a single October trade, which is unfortunate, but they’ll show whether these funds held on through a summer when Cerebras lost more than a fifth of its value and plenty of other names in the space suffered from stalling momentum.

Lam posted record sales and beat on earnings, then said supply shortages would hold back next quarter’s revenue. The stock closed at $230 Thursday, its lowest level since April. The knock-on effect is that customers want more machines than Lam can build. Management still expects chipmakers to spend more than $200 billion on equipment next year, and even with parts in short supply, it told Wall Street to expect about $8.6 billion in sales next quarter, a hair above estimates. I’ve seen a lot worse problems than that. But when expectations are off the charts, there’s no room for error.

Four of the seven funds held Lam above 1% in at least two quarters, so they know the company about as well as anyone outside the building. If Thursday’s drop was about supply rather than demand, I’d bet a few of them are already buying, but we won’t see it until the December filings land in mid-February.

Bob Byrne’s AI Stack Series:

At the time of publication, Byrne had no positions in any securities mentioned.