The Amiga did catch on everywhere but the US, pretty much. 5 million or so sold was an amazing feat at the time.
But in the US, Commodore had badly burned their dealer network in the C64 days, and they always did better outside the US and so focused their supply to Europe when they had problems meeting demand.
Unfortunately, staying in that market when it was relegated to high end niche (TV studios etc.) in the US, and largely low end in much of Europe, was a long shot. And then Commodore pursued a massively destructive campaign of hiring and firing managers in the US (which, to be fair, they had a tradition of since early in the Tramiel era) to try to find someone who could build up the US market.
But they ruined their chances further by drastically under-investing in R&D, and continuously second-guessing their engineering teams, who were likely to an extent hurt by their proximity to the toxic US part of the company.
Commodore was an odd beast in that the international subsidiaries were always mostly independent - they ordered the stock they wanted from the parent company, but decided what they wanted to sell and how. Germany also got to design some products (and there were manufacturing there). Yet most of the product development were managed strictly out of the US where their product wasn't selling well.
The A4000 was a pretty good demonstration of the consequence of the US management feuds coupled with putting engineering on the chopping block regularly: It was late; it was slow; it was expensive; it was ugly; it had IDE (the Amiga world was pretty much entirely SCSI, as IDE was CPU intensive in comparison, which was not what you wanted in a system that was already lagging in terms of raw CPU performance; this is what you got by putting managers used to PCs in charge of making decisions about Amiga projects - they saw SCSI and saw "expense" rather than one of the things that was vital to making an Amiga perform well at the time)
Before that, Dave Haynie had his "A3000+" pretty much done: Sam CPUs as the A4000; better memory bus; SCSI; space for an on-board DSP; AGA like the A4000. And fit in an A3000 case, which was actually pretty.
This was par for the course for Commodore, unfortunately. I'd like to say it was a great company, but frankly, while Tramiel did a few strokes of genius in his day, most of Commodore's history is a history of being given fantastic opportunities - often through brilliant acquisitions - and scuppering them with a gusto that would make some think it was intentional (there was/is a long lived rumour that towards the end Irving Gould and Mehdi Ali were intentionally causing share price volatility to profit of it, rather than actually trying to grow the company).
But in the US, Commodore had badly burned their dealer network in the C64 days, and they always did better outside the US and so focused their supply to Europe when they had problems meeting demand.
Unfortunately, staying in that market when it was relegated to high end niche (TV studios etc.) in the US, and largely low end in much of Europe, was a long shot. And then Commodore pursued a massively destructive campaign of hiring and firing managers in the US (which, to be fair, they had a tradition of since early in the Tramiel era) to try to find someone who could build up the US market.
But they ruined their chances further by drastically under-investing in R&D, and continuously second-guessing their engineering teams, who were likely to an extent hurt by their proximity to the toxic US part of the company.
Commodore was an odd beast in that the international subsidiaries were always mostly independent - they ordered the stock they wanted from the parent company, but decided what they wanted to sell and how. Germany also got to design some products (and there were manufacturing there). Yet most of the product development were managed strictly out of the US where their product wasn't selling well.
The A4000 was a pretty good demonstration of the consequence of the US management feuds coupled with putting engineering on the chopping block regularly: It was late; it was slow; it was expensive; it was ugly; it had IDE (the Amiga world was pretty much entirely SCSI, as IDE was CPU intensive in comparison, which was not what you wanted in a system that was already lagging in terms of raw CPU performance; this is what you got by putting managers used to PCs in charge of making decisions about Amiga projects - they saw SCSI and saw "expense" rather than one of the things that was vital to making an Amiga perform well at the time)
Before that, Dave Haynie had his "A3000+" pretty much done: Sam CPUs as the A4000; better memory bus; SCSI; space for an on-board DSP; AGA like the A4000. And fit in an A3000 case, which was actually pretty.
This was par for the course for Commodore, unfortunately. I'd like to say it was a great company, but frankly, while Tramiel did a few strokes of genius in his day, most of Commodore's history is a history of being given fantastic opportunities - often through brilliant acquisitions - and scuppering them with a gusto that would make some think it was intentional (there was/is a long lived rumour that towards the end Irving Gould and Mehdi Ali were intentionally causing share price volatility to profit of it, rather than actually trying to grow the company).