The Amiga is a very interesting case. It started out extremely capable, with a complex architecture with multiple coprocessors tightly coupled to memory and the CPU.
As time passed, that complexity made it more difficult to build newer, better Amigas that were compatible with the software already written for it. All its unique features - graphics, 2D acceleration, audio - made it complicated to improve the machine. The PC, on the other hand, was very easy - you could throw out your CGA card and get an EGA or VGA, the same way you could plug in a generic sound card and get MIDI and PCM audio, along with an IDE port for a CD-ROM.
If the value proposition held, it could have survived longer, but this business of getting a new computer in order to get better video output gets expensive very quickly. I had EGA on my 10MHz XT-compatible and it was great.
Commodore was stuck in the 8-bit home computer model - mostly non-upgradable machines. This was the 1000 (it had a little upgradability), the 500, the 600, and the 1200. The 2000, 3000, and 4000 were seen as their professional counterparts, but they were all saddled by being compatible with games written for their built-in video and audio hardware that was quickly becoming inadequate compared to PCs.
Sadly, Atari didn't have as much custom hardware it would need to evolve, but still never quite made the leap from their self-contained machines to more modular ones that'd make it easier to compete with PCs. When they did, they opted for the VME bus instead of something simpler (such as a 16-bit ISA bus).
Impressively, near their deaths, both made PCs, but the PCs didn't share anything with their proprietary jewels. Both companies should know much better.
It was only difficult because Commodore fired engineers to cut cost. They left absolute minimum staff just to keep the lights on. One example - there was no one at Commodore capable of re-designing Paula stupid PLL to accept 1Mbit clock so Amiga could support HD floppies.
They had patents https://patents.google.com/patent/US4780844A/en they had design documents for original chip, they had chip fab, they had no one with skill nor will to pay for it. All it needed was adding switchable clock dividers to control logic.
As time passed, that complexity made it more difficult to build newer, better Amigas that were compatible with the software already written for it. All its unique features - graphics, 2D acceleration, audio - made it complicated to improve the machine. The PC, on the other hand, was very easy - you could throw out your CGA card and get an EGA or VGA, the same way you could plug in a generic sound card and get MIDI and PCM audio, along with an IDE port for a CD-ROM.
If the value proposition held, it could have survived longer, but this business of getting a new computer in order to get better video output gets expensive very quickly. I had EGA on my 10MHz XT-compatible and it was great.
Commodore was stuck in the 8-bit home computer model - mostly non-upgradable machines. This was the 1000 (it had a little upgradability), the 500, the 600, and the 1200. The 2000, 3000, and 4000 were seen as their professional counterparts, but they were all saddled by being compatible with games written for their built-in video and audio hardware that was quickly becoming inadequate compared to PCs.
Sadly, Atari didn't have as much custom hardware it would need to evolve, but still never quite made the leap from their self-contained machines to more modular ones that'd make it easier to compete with PCs. When they did, they opted for the VME bus instead of something simpler (such as a 16-bit ISA bus).
Impressively, near their deaths, both made PCs, but the PCs didn't share anything with their proprietary jewels. Both companies should know much better.