Computers didn't proliferate VERY fast when software had to be, essentially, written for specific models. While 'CP/M' was a nice standard, it wasn't very standard. Apple II machines were probably the biggest single market for software (non CP/M, obviously).
The development of the IBM PC, despite its numerous flaws, with an open BIOS, published architecture and the IBM name, opened the market to (legal) clones and to business markets. The old line was 'no one ever got fired for buying IBM' in a business setting.
IBM didn't believe the PC was going to take off and be a major product - or they'd have never given the Boca Raton people who developed it the freedom to make it so open.
Without that (if IBM goes its usual proprietary route), PCs stay expensive, the software market stays terribly fragmented, and PCs don't explode like they did.
At a wild guess, some Unix variant wins out in the end, but not until memory prices come down and processor performance goes up perceptibly. Which will, in turn, take longer because the market just isn't as big.
That could easily slow things down massively. And lead to what's effectively an oligopoly of proprietary systems (Apple, IBM, probably HP and DEC, say, plus a few micros that expand out of the CP/M field.)