I suspect that they’re a very slow-moving corporate culture and likely mostly interested in value extraction over a long term. It’s a general opposite of move-fast-and-break-stuff.
You don’t expect IBM to provide paradigm-breaking sexy new things, you expect them to build boring corporate stuff that you support for decades and gradually crystallizes because you can’t break specific use cases.
It takes a specific kind to get excited about that.
The whole MCA and PS/2 fiasco is probably at its heart them overplaying their hand on a business level. The PS/2s were cleverly designed machines and MCA was impressive for a 1987 design, but they fell flat because the business guys wanted to use it to rebottle the genie that Compaq released.
I’ve heard that OS/2 got hamstrung by IBM promising too much business-wise. They sold 286s with the promise it would run OS/2, but the 286 was a pretty bad platform to juggle DOS software and more modern multitasking on, so it was jankier and more incomplete than it needed to be. Even before the era of direct competition, OS/2 had a rep of being expensive and aekward.
I wonder if it would have fared better if the MS-IBM partnership had started 3 yesrs later and targeted the 386 from day 1.