It's not too late to start buying the data center winners. Here's why

Livestream Menu Make Itselect USAINTLLivestream Search quotes, news & videos Livestream Watchlist SIGN INCreate free account Markets Business Investing Tech Politics Video Watchlist Investing Club PROLivestream Menu

It's not too late to start buying the data center winners. Here's why Published Sun, May 10 20261:15 PM EDTJim Cramer@jimcramer Late. Late. Late. That's how I felt when I saw the Roundhill Memory ETF — symbol DRAM, of course — take in more than $5 billion in a month. Super late when I saw it took in $1.1 billion on Thursday alone, all to play what the press is hailing as the memory supercycle. But then I look at its core holdings like SK Hynix, Micron , Samsung, Sandisk , Seagate , Kioxia and Western Digital , I say to myself this is a cocktail of high octane the world has never seen before. And a fantastic way to get exposure to the makers of memory all over the globe, including some not listed on American exchanges. The Korean companies are on fire. The Japanese firms are on fire. I want in. That billion that was invested on Thursday, maybe that was the smart money. Rather than late, I think that if you own this ETF you have everything that represents the incredible bottleneck of ever-rising prices where the customers can't say no thanks to the demands of AI computing. How can you say no to those kinds of investments? How do you say no to the insatiable? Late. Late. Late. I am reading my old friend Herb Greenberg's Substack newsletter, Red Flag Alerts, and what does he focus on this week? Modine Manufacturing . I know it is a smoking-hot company from its gorgeous stock chart and I know it has a tremendous reputation in heating and cooling. Modine says it has built really strong relationships with the hyperscalers. But Herb talks about how it hasn't been exactly straightforward with its transformation and how risky it's become. It does seem a little headlong. But the more I look at it, the more I think it hasn't really been discovered yet. I dig deeper and I can't believe how much these hyperscalers seem to love the company, even as Herb says that they might not even have any of these hyperscalers as clients. I'll take the other side of the trade. Modine Manufacturing is a buy. It sounds as good as Vertiv or Club name Eaton , two data center suppliers, except a little less discovered. Certainly better than Carrier , which has too much residential HVAC exposure. Plus, Modine got rid of all the slow-growing industrial. What a management team. I am grateful to Herb for tipping me off, even as the chart says I am not early. Late. Late. Late. Micron rallies over 200 points in a week without me, going from $542 a share to $747. Then again, it doesn't matter what kind of company it is, or how special its high-bandwidth memory (HBM) chips are. No company can have a stock that goes up that much in such a short period of time. It's absurd. But then I remember Micron is trading at 9 times the next 12 months worth of earnings, according to FactSet, and it may turn out to be even lower given how much the company can raise prices. So, you are buying one of the cheapest quality stocks in the universe. It's like when Melius Research's Ben Reitzes was bold enough to initiative coverage of Micron and Sandisk with buy recommendations in an April 27 note, one of the latest attempts from Wall Street to capture some memory upside. It turns out that you caught a 50% gain in Micron in just a couple weeks if you listened to Reitzes. It wasn't the latest, as in, "are you kidding me, you want to come in this late?" It was the latest, as in, "the latest, greatest way to play memory." It wasn't rash. It was conscientious. And it's just all too tantalizing. You can't read about these companies without wanting to buy them. I spent a lot of time with Director of Portfolio Analysis Jeff Marks this week, just kicking around trying to get into Micron if we just had a couple of down days. And why not? We know the demand is there. We know it's an easier way to play this supply-demand imbalance than buying the stocks of Qnity or Element Solutions . Or maybe not? Those are sleeper cell memory plays. They make the materials you must have to turn silicon wafers into chips. At least we own Qnity by virtue of fellow Club name Dupont spinning it off in the fall. Late. Late. Late. We are now buying the tertiary plays, the incidentals, because the gains that have been made already by memory, by cooling and by construction — why not Nucor ? — are way too far gone and known and we have never seen any profits like this before. Ever. But then we think of the price-to-earnings multiples. They are still so small in many, many cases. We think of things as stupid as people selling Taiwan Semi on Friday because Apple is going to do business with Intel , even as it was well-known Apple was considering using Intel's foundry in some capacity. I even asked outgoing Apple CEO Tim Cook about it a couple of weeks ago for heaven's sakes. There was nothing revelatory here.