I quote someone else saying: "The PC world might have looked very different today had Kildall's Digital Research prevailed as the operating system of choice for personal computers. DRI offered manufacturers the same low-cost licensing model which Bill Gates is today credited with inventing by sloppy journalists - only with far superior technology. DRI's roadmap showed a smooth migration to reliable multi-tasking, and in GEM, a portable graphical environment which would undoubtedly have brought the GUI to the low-cost PC desktop years before Microsoft's Windows finally emerged as a standard. But then Kildall was motivated by technical excellence, not by the need to dominate his fellow man."
That did happen. There was Concurrent DOS in the mid 80s, which was multitasking DOS. GEM was used in several other computers, but also was a competitor to Windows in the early days - because it was extremely lightweight and you didn't need to upgrade your 8088 PCs to run it.
But low cost licensing they didn't have.
In the famous "IBM visits Digital Research", the narrative often heard is that IBM came to Digital Research (Kildall's house) and his wife said he was out flying. That was all true. However, IBM did not leave nor demand Kildall return - the reason is that Kildall's wife handled the business side of Digital Research while Kildall did the technical side. Thus, he wasn't needed for the meeting at all.
The main sticking points were that IBM wanted an NDA signed, and Kildall's wife refused. The second demand was that IBM wanted an all-in price to license CP/M. They didn't want to have a per-PC license, they wanted an all-in price.That was the biggest stumbling block because DR did not offer an unlimited seat license.
So after that they went back to Microsoft to get an OS from them (Microsoft was tasked with producing and porting their suite of languages to the new computer). Microsoft did cheat Seattle Computer Products when it bought QDOS from them - because they wanted to pay per-customer. And they sold it to well, one customer - IBM. So that's why PC-DOS and MS-DOS existed and why IBM kept shipping PC-DOS. They bought it from Microsoft and had Microsoft maintain it for a while.
In the end, Digital Research did sue IBM and Microsoft because MS-DOS/PC-DOS looked too similar to CP/M (one of the first "look and feel" lawsuits - DR was known to sue companies for making something look like CP/M). IBM settled out of court by offering to ship their PCs with CP/M in the end. (The PC never shipped with an OS - that was something the IBM reseller added as a package deal). Of course, CP/M couldn't compete in the 16-bit world because MS-DOS was $99, while CP/M was $250 and never really caught on, because by that time, people have ported their CP/M applications to MS-DOS/PC-DOS. (Of course, Microsoft had a hand here - they made some source translator tools - thanks to the similar architectures of the 808x (8080, 8085, 8086), 6800, 650x and Z80 CPUs, it was possible to do a mechanical source translation to aid porting. MS-DOS/PC-DOS/QDOS was structured after CP/M's design which helped greatly because system calls ended up being similar.
I don't know how much CP/M cost on 8-bit computers - it could be cheap, it could be expensive. Just on the PC side, MS-DOS was far cheaper than CP/M ever was.
CP/M produced MP/M which was a multi-user version of CP/M, which begat 16-bit versions CP/M-86 and MP/M-86, which then became Concurrent CP/M-86, which evolved into Concurrent DOS. That eventually transformed into Multiuser DOS in the 90s (after Novell acquired Digital Research).
Incidentally, MS-DOS 1.x/PC-DOS 1.x was very CP/M like. MS-DOS 2.x started adding more traditional operating system conventions that Microsoft adopted from their Xenix. In that in CP/M, and DOS 1.x, files specified on the command line were opened for you by the operating system and you basically manipulated pointers and asked the OS to bring it into memory. DOS 2.x later acquired the more traditional open/read/write/seek/close style semantics and system calls.