You can always tell when a company does not have a stable succession plan in place after a CEO transition. The change happens suddenly, and it doesn’t seem like anyone was expecting it. Lower-ranking people in the company get the call up to the majors, but because they don’t know the entire landscape, multiple people are brought in to be CEO. The speculation about what happened is kicked into overdrive.

And there’s a vague promise that there will be an in-depth search for the replacement CEO.

This happened earlier this year with Paramount Global, when three executives—one who came up in traditional broadcast, one who came up through MTV, and the third best known as the director of Good Burger—replaced Bob Bakish, who was let go amid messy merger talks that led to a lot of historic content being removed from the internet.

Now, it’s happening with Intel, albeit in more dramatic fashion. On Monday morning, the company announced the retirement of Pat Gelsinger, a legendary figure within the company who played a key role in designing the blockbuster 486 chip—and word is that, if he hadn’t retired, he would have been forced out. He loomed large, and approached his CEO role at Intel with ambition. His vision of the modern Intel was as a company that did not just produce its own chips. but fabricated designs for other companies. This is the position that TSMC has embraced, with its capabilities giving companies like Apple and AMD the ability to leapfrog Intel’s chips on a technical basis.

But resetting a legacy business around a different paradigm is difficult, especially if that business makes things. And trying to catch up was so difficult that it was leading to insane losses—$7 billion in 2023 alone, up from $5.2 billion in 2022. Had Intel not invested all this money into trying to compete with TSMC, it would have been profitable. (The headaches of doing so were so serious that Intel had to outsource some of its chips to TSMC anyway.)

The result is that, unlike many of its competitors, which have been able to move nimbly because they are only making the diagrams for the widgets, the company has missed out on some important trends in the chip industry. It didn’t have much of an answer to the rise of AI, for one thing, and that created an opportunity for a new competitor in laptop CPUs, Qualcomm, earlier this year.

Intel, one of our most important companies, has looked less important by the day—to the point where, in September, the idea of Qualcomm acquiring the legacy giant was floated by the friggin’ Wall Street Journal. (The pitch was serious, but interest is reportedly fading on Qualcomm’s end due to the sheer complexity of Intel’s business.)

Intel’s problem is that it wants to fabricate chips on contract for other companies, and that’s an extremely expensive business to build out. (Miles B./Flickr)

Given its American roots and the odds that more than half of you are reading this on a computer with an Intel chip, it matters to the economy that Intel, the developer of many of our modern computing standards, is strong. This is highlighted by the fact that the Biden administration granted the company $7.86 billion to build U.S. manufacturing facilities just this week—a lot of money, albeit a slight haircut on what the government had promised them. But it’s simply not particularly strong, and solving that may not be easy.

You could point to many reasons why Intel is struggling. For one, its lack of innovation in chip design opened a lane for ARM chips in both server rooms and embedded systems. For another, it has been winnowing down some secondary playing fields in recent years, some of which showed significant promise. It quit building its own NUCs just as the mini PC market was starting to blow up; it also pushed aside its Ethernet switch and data center design businesses. It has cut investments in R&D in some areas—giving up an important innovation pipeline that, over the years, has led to major success stories like PCI and USB.

By cutting investments in areas that could lead to future success stories, Intel has essentially bet that the fab path is the way to go. In this light, Gelsinger’s bet looked increasingly risky, and the market responded accordingly.

It’s hard not to look at the decline of Intel, a company whose success created the PC industry more fundamentally than any other company save Microsoft, and not wonder if we have reached a point of diminishing returns. After all, it’s not like there haven’t been companies in prior tech eras that looked unimpeachable only to falter entirely and fade into history—to name a few, Digital Equipment, Silicon Graphics, and Sun Microsystems. Even IBM, as stable a name as you’re able to find in the tech space, is not what it was.

Intel is at the center of every device we own. It was once too big to fail. Yet, as the Gelsinger situation shows, it is not immune to toppling over, allowing other companies to pick up the pieces as they try to speedrun an AI fever dream. Companies of Intel’s size don’t go away. They gradually get subsumed into their competitors until there’s nothing left.

“Intel inside” used to mean something.