Naughty Flicks

Skip NavigationMarketsBusinessInvestingTechPolitics & PolicyVideoWatchlistInvesting ClubPROLivestreamMenuWe don’t want to get caught too deep in the weeds. Don’t overcomplicate things. It will only lead to suboptimal thinking. I start with that simple precept because when you’re facing the biggest challenge yet to The Greatest Story Ever Told — the rise of the data center — you just need to follow the money, as the reporters did in the Watergate investigation. You’ll find that the investment story has broken down, and the money has moved on — perhaps not for good, but until the dust settles around the data center buildout. Let’s revisit the brutal tableau of last week. We had a debilitating sell-off, a real “give up” that was unlike anything we have seen. That includes the collapse in late July of Situational Awareness , that ridiculously hyped hedge fund led by Leopold Aschenbrenner that blew up because it had no risk controls. I have contempt for that kind of recklessness. Throughout my career, I’ve prioritized risk control over stock picking, dating back to my time at Goldman Sachs in the 1980s. If you don’t have risk controls, get the heck out of the business. But Situational didn’t get blown out because of Aschenbrenner’s thesis: that there may never be a better investment story than the data center buildout, which touches so many sectors. It got blown out by a momentary glitch, magnified by his lack of controls. This time, it was the thesis itself that blew up. At the exact same time, we finally got a countervailing thesis: the prospect of lower interest rates. Rates always trump any other concept in investing because they are the sum total of the forces shaping the market — the Federal Reserve, the economy, inflation, Treasury debt, money demand. Rates are too big to lie. Rates fell slightly last week, and all I heard about was a “flight to quality” among investors. But that’s only partially true. We are beginning to see a slowdown in inflation. That quarter-point rate increase in September is proving very powerful when coupled with a determined Fed Chair in Kevin Warsh, who doesn’t seem to be playing ball with the president. Warsh seems independent, and people didn’t count on that. The shift is real, and so are the alternatives beyond the AI trade — food, drugs, and soft-goods stocks. Biotech may offer the best long-term opportunities. They have been transformed from fly-by-night fireflies to the next Merck or Eli Lilly . AstraZeneca , a perennial second-rater, gets rerated in real time. These alternatives to the AI trade might have staying power. Maybe the valuations of the winners keep growing? Can we pay $200 for Merck? Maybe. Seems excessive. But didn’t we start paying incredibly high multiples for tech? Isn’t that a reason to take some profits in Apple unless you are as storied about holding it through thick and thin as I am? (Obviously, I am signaling that if you want to step aside from the impossible situation of Apple at 40 times earnings and revisit it, that makes sense.) Do we just buy into the rotation? The answer to that depends on your investing strategy. If you share ours, it gives you an automatic buy-in, hence the endless references to Johnson & Johnson . Hence the Kimberly-Kenvue gambit. Hence the love of Coca-Cola . The AI buildout remains the elephant in the room. Lower rates may make alternatives more attractive, but they don’t make The Greatest Story go away. What’s really happening? With the midterm elections looming, the moment has become so fraught that even the diehards are struggling to bear the pressure. Let’s go over the three big forces that are defeating the data center. First, the election is allegedly a referendum on the data center. We know they will still be built. But if the Democrats do well, we have to presume fewer will go ahead. For example, if we need 200, only 150 are put in motion. Isn’t that why GE Vernova is down? The gas turbine maker reported zero cancellations and an order book that continues to grow, even though its quarter wasn’t perfect . Investors bailed anyway, compounding a dreadful week capped off by the gang that couldn’t shoot straight at Intel . The hyperscalers and the bitcoin managers turned data center builders could suffer if the Dems score a landslide. Lawsuits make it a disorderly election unlike any we have seen. We know data center builders didn’t hire enough people to operate their facilities once they were finished. Initially, they despoiled the water supplies and raised electricity prices. That’s no longer true. Data center builders subsidize where they build, not the other way around. However, the damage was already done, compounded by the endless discussion of jobs taken, led by none other than the financially suicidal Dario Amodei. The Anthropic CEO had the temerity to say so many self-destructive things about the coming dinosaur age of white-collar jobs. When you hear AI could displace half of all entry-level white-collar jobs in the next one to five years, as Amodei said, you can see the dinosaurs trudging to their deaths. He and OpenAI CEO Sam Altman are such quizzical figures: so dreadfully immature about the power of their words, so cavalier about the consequences. How do they know jobs will disappear? Why can’t they see AI’s positives — even recursive systems, where gremlins create more gremlins — could create new jobs and maybe more of them? Why be so downbeat that it gives ammo to AI’s biggest haters? Can’t they just say, “Some good, some bad?” Can’t they at least relate their own work to previous industrial revolutions, giving a nod to the optimistic yet historical Nvidia CEO Jensen Huang? Why embrace the hysterical over the historical? So we are supposed to side with them? In the end, could the AI leaders’ own hubris — all this much ado about nothing — kill Anthropic’s IPO? That would be worse than the electorate rejecting the AI buildout itself. That wouldn’t be a nail in the coffin; it would be cremation. What’s the reality? There will be plenty of data centers built either way. But the big move from December, when we realized how gigantic the build-out could be, still seems destined to be continually reversed. Does it go back to December? The country’s in a mood to turn negative on the AI players. So we have more bans and more siting issues, and we have to take numbers down no matter what. That was the proximate cause of last week’s debacle. Second? Nvidia. As you know, Nvidia has been my favorite stock for ages. It makes the most money. It is the sun around which all the data center planets revolve. But Nvidia is starting to look like what Richard Nixon called the United States in 1970: a “pitiful, helpless giant.” Nvidia is the pitiful, helpless giant, now revealed, according to the bears, to have just propped up an ecosystem for its own earnings, with a CEO now revealed to be a dubious character. Now your author, showing fealty to the man and his company, completely disagrees. According to a report in The New York Times last week, Huang formed a close relationship with a top executive at a Singapore-based data center company who was found to be a spy passing information to Beijing. Huang can’t stop a customer that appears to need every chip it can get from transshipping some to China. That’s been happening for decades. The implication is that President Donald Trump was informed of the situation but looked the other way because Huang flattered him — a view even people outside the stock world now hold, thanks to the Times’ story. In truth, all executives, every one, play the same game with the president. Trump has identified the winners and he wants to help them. Jensen’s just doing what Jensen has always done: selling chips to help his company make money and to ensure the U.S. dominates data center culture. He wanted to sell his chips to China again because he knew it was good for shareholders and for the U.S. if China didn’t develop its own ecosystem and instead relied on Nvidia’s. There isn’t more to it. But I also know that my view was irrelevant. The writing on the wall says the reverse continues, and the pirouette rally-to-intraday crash was the top. It’s not too late to sell, but to buy it back when it’s right again is very hard. But I concede that the stock is untenable for almost all right now. Why is that? Why have the long knives just gotten sharper? Simple: the hatred for the data center hasn’t just reached the electorate. Everyone who is a part of it wants out. We are sick of the neo-clouds and their inability to pull off what they said they could. We are dubious about their balance sheets. We know their bonds suck. We know their orders are suspect. Most important, Microsoft is the best stock in this godforsaken group because it trimmed capex just enough to stay within its means. It isn’t issuing shares through at-the-market sales that could put a lid on the stock. Its core business isn’t eroding, as at Google. And it isn’t promising that returns are just around the corner, as Amazon is. Amazon’s results echoed Gertrude Stein’s line in “Sacred Emily”: a miss is a miss is a miss. Microsoft is now praised for being reasonable with its spending. That’s not what you want to hear if you own Micron , Iren or Nebius . The latter two are vulnerable and are only propped up by the ridiculous amount of short interest. Amazon, though, gets chastised for trying to meet its customers’ demands. CEO Andy Jassy and CFO Brian Olsavsky are willing to spend beyond their means because they see a moment when they could win. They see that Alphabet’s core business, aside from Google Cloud, has become a mess because it seems to be set up now for its advertisers, not its customers. And it’s happening at the very moment when its AI assistant Gemini has dropped back below ChatGPT, Claude, and even Grok. Jassy seems to recognize that Google is hobbled — something that makes me want to sell our position. We’ve held on because of its lower valuation and my belief that its leaders aren’t chowderheads, oblivious to how serious the threat is. Jassy knows that Microsoft is getting credit now, but won’t two years from now when Amazon Web Services is still growing at 40%, and its retail business is closing in on a trillion dollars. He is in the driver’s seat, even if the spending is momentarily punishing shareholders. Unlike Intel, which investors now despise, Amazon said there will be a huge return on its spending beginning in the second half of 2027, when its semiconductor business takes off for massive profitability. Believe Jassy at your own peril? Or let ‘er ride? I am in the latter camp, which comes off right now as a death wish, but I am not a hedge fund manager, and I won’t be judged by short-term thinking. And lastly, there are the stocks themselves. I’ll tease you with a few thoughts now and save the in-depth analysis for Friday’s Annual Meeting. Suffice it to say, nearly all of the quarterly results have been excellent, with one exception: Kioxia , the Japanese memory chipmaker, lowered its guidance. That doesn’t just show that Kioxia is a mess, but also that the group is unreliable despite SK Hynix’s extraordinary numbers. Sure, GE Vernova ruined the short term, and yes, Caterpillar became more of a truck company, and we didn’t get the order spike from Lam , but earnings aren’t reverting to pre-2025 levels. And the stocks shouldn’t revert to December of 2025, before they levitated. Yet, should doesn’t matter. At this very moment, we are looking at the denouement and trying to figure out whether it will come after the election and the president’s move to send oil to the $150s. We are trying to figure out if Marvell Tech ‘s bid for Lumentum actually happens and whether it sends both stocks up or down, the latter being dreadful and giving us nowhere to run, nowhere to hide. So I leave you with the mindset of every long manager except the braindead or the brainwashed, and I at least have given you how people feel, with the best possibility being that the election is a clearing event and the neoclouds don’t collapse and crumble regardless of the outcome. (See here for a full list of the stocks in Jim Cramer’s Charitable Trust.) As a subscriber to the CNBC Investing Club with Jim Cramer, you will receive a trade alert before Jim makes a trade. Jim waits 45 minutes after sending a trade alert before buying or selling a stock in his charitable trust’s portfolio. If Jim has talked about a stock on CNBC TV, he waits 72 hours after issuing the trade alert before executing the trade. THE ABOVE INVESTING CLUB INFORMATION IS SUBJECT TO OUR TERMS AND CONDITIONS AND PRIVACY POLICY , TOGETHER WITH OUR DISCLAIMER . NO FIDUCIARY OBLIGATION OR DUTY EXISTS, OR IS CREATED, BY VIRTUE OF YOUR RECEIPT OF ANY INFORMATION PROVIDED IN CONNECTION WITH THE INVESTING CLUB. NO SPECIFIC OUTCOME OR PROFIT IS GUARANTEED.Read More