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…
Read More: Here’s the tech stock to buy in a market losing patience with tech