I think that a good part of that killed Amiga and Atari was the fact that IBM PC were the "serious" computers, the one you bought for work, and the Amiga and Atari were less considered. I know that they had some success in their niche like music and publishing for Atari, and graphism and design for amiga, I even saw Atari on a warship. But at the end most of the money at the time was made in a professional setting, and neither Atari or Amiga managed to get a sizable part of it.
I regret it because passing from GEM to DOS really felt a step backward
There were two reasons the IBM PC clones became dominant. The smaller reason is because IBM created it. The bigger reason is because everything but the BIOS was off-the-shelf parts and easy to copy. Apple had the same problem with Apple II clones where only a couple of companies managed to not lose when sued by Apple. IBM couldn't sue Compaq and others for their clean room BIOS replacements.
If the IBM PC had custom chips, it would not have been easy, or even possible, to clone and would have remained an IBM-only product. If that had been the case, I suspect CP/M and GEM would have won out and Microsoft would be much smaller and still producing language compilers today. The only question is, which version of CP/M would have won out? I think Motorola's 68K series would have won over Intel's x86 series.
The problem with motorola winning the 80s was that no matter what they did, they were facing the same problem in the 90s, the 88k was going to flop hard. And early PowerPC would still have been 'so so' compared to the Pentium and onwards.
That said, would the Pentium have happened without the PC industry - everything up to the 486 would likely have continued to develop as it did, even without the PC as they were strong embedded chips too.
I remember reading magazines when the Amiga 1000 was introduced, and even then the general conclusion was: amazing hardware, but doesn’t run PC software.
I regret it because passing from GEM to DOS really felt a step backward