IBM had made big commitments to enterprise customers to get them to adopt the PC's successor, the PC/AT (https://en.wikipedia.org/wiki/IBM_Personal_Computer/AT

IBM had made big commitments to enterprise customers to get them to adopt the PC's successor, the PC/AT (https://en.wikipedia.org/wiki/IBM_Personal_Computer/AT https://en.wikipedia.org/wiki/IBM_Personal_Computer/AT), which used the older, transitional 80286 processor. These commitments created a big drag on lots of other IBM projects, as they couldn't do anything that would be perceived by those enterprise customers as devaluing their investment in the AT. So not only was IBM slow to adopt the vastly superior 386, it also insisted on things like OS/2 having to be compatible with the 286, where Microsoft was free to let their competing product, Windows NT, abandon the 286 and all its baggage completely. This is a great illustration of the idea of the "strategy tax" (see http://scripting.com/davenet/2001/04/30/strategyTax.html http://scripting.com/davenet/2001/04/30/strategyTax.html), where various parts of a business that are deemed non-strategic have to give up things that would make them more competitive to prop up another part of the business that is deemed strategic. IBM's bet on the 286 was so big that they were still pushing 286 machines well into the PS/2 era, by which time the clone makers had long since flocked to to the much more capable 386. Those commitments to enterprise customers meant that the 286 became an anchor tied around the waist of IBM's PC business.