Actually IBM stuffed up with the PC.
It was originally to be marketed to hackers to fill a small niche market of those who were into microcomputers, like many of us here. But quickly they realised the product had wider merit because it became the defacto standard for hackers and spread into the general corporate world. They had no vision of how it would change the much of the world.
Er, no. The PC for a niche market was the one before the IBM PC we all recognise today. It had a proprietary IBM CPU and you could only use an interpreter on it (some kind of BASIC, I think). The IBM PC we all know was a result of IBM worrying about the number of small businesses who found they could do their basic accounts and database activities on the Z80 based CP/M machines which were becoming popular. IBM made a sort of "next generation" CP/M machine, with a 16 bit core. They wanted a 16 bit CP/M on it, but when Digital Research didn't play ball, they got a CP/M clone from Microsoft - MSDOS.
They used the Intel 8088 micro - a segmented addressing nightmare and crappy architecture in comparison to the superior Motorola 68000.
The screwup here was Motorola's. The 68008 was so late it was irrelevant when it appeared. There was no way a full 16 bit bus or the huge expensive 64 pin ceramic package of the 68000 was going into this simple machine. If the 68008 had been properly planned into the release of the 68000, in a low cost 40 pin package, the IBM PC would probably have used it.
The biggest screw up is that IBM published the complete BIOS listings and the circuit diagrams, so that other companies could make option cards like graphics cards etc. IBM made a gross miscalculation. Instead the more lucrative market was in cloning the IBMPC. Even though most of them were crap quality, the prices of IBM Clones were a fraction of the IBM machines.
That was actually a very smart move by IBM. They didn't want to stop competitors. Anti-trust concerns meant they only wanted to slow their competitors down. They did enough to ensure that, while making the machine open enough that third party cards flourished, and the PC ended up in all sorts of unexpected applications, broadening its market.
IBM got greedy for money with big fees on Microchannel architecture and the overpriced OS/2 Warp operating system.
Microchannel was a very stupid idea, but OS/2 Warp wasn't around in those days. The original OS/2 was a screwup, mostly because it was locked to IBM's microchannel hardware. That is what seriously limited its appeal. I don't remember it being very expensive. The whole machine was more expensive, as a full scale OS like OS/2 needed a lot more memory than was being fitted to most machines at that time.
If they had one ounce of innovative marketing skills, they would have given every university student on the planet a fee copy of OS/2 Warp, and Microsoft would have amounted to not much. But they paid for their ineptitude because now OS/2 is dead and buried. The funny thing is IBM thought the world would accept OS/2 if all the employees were forced to use it. In fact IBM employees were forced to use OS/2 ("Orrible System 2") because they thought the religion would spread. In the days of the dodgy Louis Gerstner, IBM employees were forced to use OS/2, when the real world was embracing Windows. Incidentally Microsoft wrote OS/2, not IBM. And only 16 bit applications could run on OS/2's pathetic Windows shell, not 32 bits.
OS/2 was IBM's design and product. Microsoft were just body shop contractors helping the pool of IBM developers. At that time there were only 16 bit Windows apps, so I am not sure what your last sentence is supposed to mean.
OS/2 had a lot of good qualities, but its key problem was it was seen as an attempt by IBM to lock people in. Most of us thought "the only reason they aren't putting Unix on that thing is because they want to screw around with us" and we mostly ignored it. Strangely, when Microsoft launched NT some time later, and there really were 32 bit Windows apps, many people had a very different attitude. They saw it as the dawn of low cost versions of things which had always run on Unix, like super expensive CAD tools. They had no justification for why that massive price reduction should occur, when it changed nothing in the cost structure of the applications supplier.