It’s not the spend, it’s the return. That’s what Friday’s technology stock plunge said. I think it just might be the most significant tech selloff in more than a year. We need to know — potentially — what we are facing. We need to know if the trillion-dollar spigot is drying up — or it is just a dry spell. For my Charitable Trust, the portfolio we use for the CNBC Investing Club, we have been consolidating and gradually trying to shrink traditional tech — semis, software, data center — and move into other kinds of tech, namely tech-infused pharma and aerospace. We tried to make Intel — not Nvidia — the focal point of the portfolio, concerned that there are not enough new move-the-needle customers still out there for Nvidia. No, I’m not giving up Nvidia. It is still amazing, and I think it will have a bang-up quarter. But the “action” in the stock is speaking too loudly. The action in Apple is screaming that its decision, made intentionally or de facto, not to spend hundreds of billions on AI, is brilliant. It’s having the best month in three years. Tons of critics second-guess Apple’s decision-making. That’s wrong. Apple decided a long time ago that much would flow to it if it made the best handhelds. It does. That allowed it to pick and choose which hyperscaler-chat-bot company it wanted to affiliate with, because they have quickly turned into commodities. Google had no choice but to virtually give it away — at least on a net basis — because Google Search has become suspect in its return while Gemini is no Claude from Anthropic. So, we gravitated to a new company, late for now, Intel, because we could see that the ratio of graphics processing units (GPUs) — Nvidia’s giant, expensive chips — to central processing units (CPUs) — Intel and Advanced Micro Devices (AMD) — and perhaps, Arm Holdings , if it can get foundry time, even as it is partners with Intel) was quickly changing. When Lip-Bu Tan took over as CEO of Intel, the ratio was about four GPUs for every one CPU. Now, he told me last Thursday, it’s about one CPU for every one GPU. Soon, data centers will have four CPUs for every one GPU. The gross margins on the GPUs are far more bountiful than on CPUs. A well-run Intel can change that. This is a well-run Intel. Plus, the CEO is perhaps the most dedicated semiconductor investor who knows how to spend the money wisely to build foundries (factories to manufacture chips) that are in short supply. Most important, he knows packaging, which is the equivalent of bundling CEOs to make them more powerful now that it is getting harder and harder to make nodes smaller and more powerful. The notion of Moore’s Law — Gordon Moore, an Intel founder — that you can keep making ever-more-powerful smaller chips may have run out — Or definitely has run out, according to Jensen Huang. When in doubt, go with Jensen. So, we went with Intel, betting on an upside surprise. As usual, we buy slowly for the Trust. We had a little more than half a position on betting against ourselves that the company would report a spectacular number. We lost the bet. It was spectacular, maybe better than that, and it traded up more than 10% in last Thursday’s after-hours session. High-fives all around. We looked good. Now I was out on Long Island to throw a rehearsal dinner for special nuptials, the marriage of my stepson, Will Detwiler, who has had that position for 21 years and lots of rearing, and his incredible now-wife Caroline (Win With) Willkie. It was a rather sizable prelude to the actual affair, a rehearsal dinner so large that it had a rehearsal for the rehearsal dinner. I was confident of the market’s reaction to Intel even as I lacked confidence in the market. It’s more than a tad difficult being long much of anything when you have a president talking about saturation bombing of a crafty opponent that seems to be a state that has more missiles than people. We’re back to where the war or wars or who knows what can drive oil to where bears can confidently discuss 5% to 6% inflation. My confidence was misplaced. Entirely. I got up early Friday, not to disturb the participants, took a call from American Express , which we told you would cause selling pressure, and out of one retina saw Intel trading not at $109 but $106.30. Then, $106.28 and then $106.20, and the cadence was both sickening and relentless. Hardly a stand was made before market hours. Count the upticks on a couple of hands. After it went to $103 and change, soon before Friday’s open, you could calculate how much it would be down on the day. I switched mentalities and went into thank-my-lucky-stars mode to have plenty of room to buy; if we still wanted to. Which brings me to the essence of this beautiful Sunday’s piece after a picture-perfect wedding where I forced myself to be in the non-stop present, something I had only been capable of at one wedding-mine — and two Super Bowls. So, what the heck happened here? Why did Intel close down nearly 8% on Friday? I have heard lots of reasons for the tech sell-off. The most preeminent is that the market has decided to exercise its power to stop the spend. With Alphabet ‘s stock reeling almost forty points below what once looked like a terrifically priced secondary — at least, before we heard that Google was going to up its capex again — it looks like many sellers decided enough is enough. We won’t reward spending with a higher market capitalization. That was a brutal judgment, especially given that Google Cloud had an implausibly tremendous quarter — cue the interview of Thomas Kurian, the head of Google Cloud and often called the LeBron James of tech, a statement with refreshed poignancy given my proclivity for the Philadelphia 76ers. That Alphabet call had me wishing for Ruth Porat back in the CFO role, and her 80-plus calls under her belt to explain why the spend keeps getting upped. Yes, we can say “to meet demand,” as now CFO Anat Ashkenazi said multiple times while sprinkling in positives about Alphabet’s balance sheet. I have found both statements painful. That’s because Alphabet’s balance sheet, with an income statement showing negative cash flow, is no longer the belle of the ball. It’s okay, not great. If you owned Alphabet for its rock-solid balance sheet and buyback, you are thinking of a different Alphabet, one of yesteryear. The real pain, though, comes from this “meet demand” statement that we heard so much. I am not sure if Alphabet knows what it means to a market that’s starting to lose a lot of money in these hyperscalers when it hears “meet demand.” We are definitely not looking for companies to “meet demand.” We are looking for them to “make money.” We aren’t hearing anything like that. The only company making money here with Alphabet is Apple. This “meeting demand” stuff is wearying. I feel like the Trust owns companies that are losing fortunes on everything they make, but they are going to make it up in volume. What’s happened is, at last, we have come to accept not that these companies don’t know what they are doing. I am sure that Alphabet feels that if you build it, profits will come — but there are better stories elsewhere. What do I need this spending horror show when I can get behind Club name Johnson & Johnson ? The technology behind the materials science of 3M doesn’t require billions of dollars to lose, well, billions; it just makes money. Again, I am not in the camp that the emperor has no clothes. One of these emperors to be will, perhaps, Anthropic because it is business-to-business (B2B) and we love the stickiness of B2B; the fickle nature of the OpenAI business-to-consumer (B2C) paradigm is viewed to be the culprit for the leaky bucket of that shop. Oh boy, OpenAI has to come public in the worst way; thank you very little. In some ways, it doesn’t matter whether you sell tech because you don’t see a return or you sell tech because you think these companies lack any discipline. Either way, hundreds of billions of dollars are leaving the cohort. That’s why I am wondering if this damned-the-sellers-full-speed-ahead mentality is coming to an end. Which brings me back to the all-encompassing selling of Friday. When you are watching the stock of Intel sink, you can say that it’s going down because you didn’t get the Wall Street analyst price target bumps you thought it would. They were disappointing. But AMD had a meeting simultaneously, and the only companies that can give Intel a run for the money right now are Nvidia, when it reports late next month, and AMD right now. Shares of AMD sank, too, but not as viciously as Intel, but at a brisk pace. The selling was ferocious. By the end of the session, as the wedding flowers abounded, the Fosforo (my wife’s agave spirits company) flowed, and the toasts were readied, you could almost see the selling morph from being about the discipline of capex to a rumor of a cut in capex by a hyperscaler. That someone blinked. I don’t know if someone actually did blink. I suspect we will find out Monday. But if someone, other than a supplier, doesn’t come out soon and say “we are beating and raising our numbers,” I suspect that a trillion dollars’ worth of market cap comes right off the top of all the behemoths who still ply AI. No illusions here. We cannot have a sustained rally with tech bleeding through the eyeballs. There are too many companies that have run up to see that happen. We know that the Dow Jones Industrial Average has been able to resist, and the S & P 500 has only slightly faltered, but the Nasdaq is back in charnel house mode. I say we keep tilting away from traditional tech. We stay close to see if someone forecasts a profit, which could lift all boats, and we recognize that there are not enough companies left who still need the hardware, at least not right now. Bottom line Why, then, Intel? Because the only way to amortize all of the spend now is with AI agents, and agents are run on Intel’s CPUs. That and robots. We know that robots are such a huge market that they can’t be left just to Tesla. I think all of the hyperscalers are going to have to offer them. They are not a GPU product as much as a CPU market — robots are packed with them. Right now, we have lots of use cases both in and out of the data centers, but the next great demand wave, I believe, comes from robots that are B2B to business and then B2C. The current demand for all sorts of tech will sustain the CPU. But I think that robots give us the use case that will keep the stocks in some demand after the companies own up to what might be a sensible pause in spend. I sense we will know soon enough whether it’s just a rumor of a pause or an actual one. Either way, keep moving toward new and different tech while backing only Intel when it comes to old tech. A pause could be so jarring that, initially, we won’t buy the pausers, but let’s not get ahead of ourselves. And remember, a beat and raise from just one hyperscaler changes the entire equation; and three more of them — Amazon , Meta Platforms , and Microsoft — report earnings this week. And oh yeah, fellow Club holding Apple also reports this week . (Jim Cramer’s Charitable Trust is long INTC, NVDA, AAPL, AMZN, META, MSFT. See here for a full list of the stocks.) 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.
Here’s the tech stock to buy in a market losing patience with tech
Jul 27, 2026
