IBM is still IBM.

In procurement they say "No one gets fired for buying IBM" but idk...

The Saying and What It Actually Means

"No one ever got fired for buying IBM" is technically still true, not because IBM is the right choice, but because everyone who would fire you also bought IBM, and firing you would invite scrutiny of their own decisions. It is not a statement about quality. It is a statement about organizational risk distribution. The IBM contract survives the people who signed it. The implementation team that failed to deliver does not. The saying has a missing second sentence: plenty of people have gotten fired for the IBM implementation going sideways. IBM remains engaged. The humans do not.

The Pivot History

IBM has been the same company in the sense that it has always been large, always had government contracts, and always found a way to make money on whatever technology was currently important. The specific technology it was important in has changed continuously. Tabulating machines gave way to mainframes, which produced the PC, which IBM invented and then immediately lost to Microsoft and Intel when it chose an open architecture. Services became the story under Lou Gerstner in 1993, which is where the current IBM really starts — a company that realized the hardware margins were gone and pivoted to selling expertise and process.

Watson was the AI pivot, announced after the Jeopardy win in 2011. Watson was going to cure cancer, transform healthcare, reinvent financial services. Watson Health was sold in 2022. Watson is now watsonx, which is the current AI pivot, which is the same pivot described with a different product name. The Red Hat acquisition in 2019 for $34 billion was the hybrid cloud pivot, and it was probably the most coherent strategic move in two decades — Red Hat had real products, real customers, and a community that was not going anywhere. What IBM does with that community over the next decade will be the actual test.

The Mainframe That Will Not Die

The IBM Z mainframe is a genuine engineering achievement. It processes a significant fraction of the world's financial transactions. It has extremely high reliability. It has been "about to be replaced by commodity hardware" for approximately forty consecutive years and has not been replaced. The people who understand it deeply are retiring. They are not being replaced, because the incentive to develop deep mainframe expertise has diminished. This is a different kind of infrastructure problem than the ones that get written about in ops blogs, and IBM is not obviously equipped to solve it, but neither is anyone else. The mainframe will still be running when someone finally figures out what to do about it.

zLinux runs Linux workloads on mainframe hardware. It was genuinely ahead of its time when it was announced. "zLinux is going to be big this year" has been said for roughly twenty consecutive years. It is not big. It exists, which is more than can be said for most things IBM has announced in the same period.