When was the first 68040 accelerator for Amiga on sale?

[Bruce Abbott]

Originally Posted by Thomas Richter

Maybe not "most of you", but "most of some other group of customers" that would have had the money to sustain the development of the platform, other than "most of you".

I was one of those 'other customers' who had the money to afford an A3000 at full price. I knew local people with A2000s and A2620 or A2630 cards too. There weren't enough of us.

According to Dr. Peter Kittel, German Amiga sales consisted of 1499800 'low-end' units, 27,500 'mid-range' (A1000), and 150180 'high-end' (generously including 124000 A2000s). So that means 9% of Amigas sold in Germany could be considered 'high-end'.

Germany was quite exceptional in that they sold ~6000 A3000T's, ~3.6% of the total. Unfortunately we don't have a breakdown for sales in other countries, but I'm betting very few truly high-end models were sold elsewhere. Also many A3000s were sold at less than cost price to move them before the A4000 came out.

It was foolish to expect less than 10% of your customers to carry the company. The people with the money to afford eg. an A3000 could also afford a high-end PC or Mac, or even a UNIX workstation. The only reasons they wouldn't choose a different platform is that they were hardcore Amiga fans, or needed an Amiga for some specific use case (eg. video editing).

I understand that Commodore did sell a very small number of A3000UX's to universities, but overall it was a failure. If only they had put that effort into developing AA for low and mid-range models instead (to be released in 1990 in place of the A3000) they might have made enough money to continue development into the mid 90's. But the engineers were too set on getting Unix onto the Amiga, to satisfy a tiny proportion of their customers (and themselves).

BTW this focus was not new. Commodore started developing the C900 series (with Zilog Z8000 CPU) in 1983 under Jack Tramiel's leadership. This would run a UNIX-like OS compatible with System V. The workstation model had a 1024 × 800 display resolution, while terminals had 80 × 25 color text or 600 × 400 color graphics using the chip that was eventually put in the C128. This was very high-end stuff that only businesses would buy. In the end of course it failed, luckily before being produced in quantity - another unnecessary drain on Commodore's R&D resources.

[TCD]

Originally Posted by Bruce Abbott

It was foolish to expect less than 10% of your customers to carry the company.

They don't need to 'carry the company', they just need to pay for your R&D.

[Thomas Richter]

Originally Posted by Bruce Abbott

I was one of those 'other customers' who had the money to afford an A3000 at full price. I knew local people with A2000s and A2620 or A2630 cards too. There weren't enough of us.

You brought this argument before, and it is as wrong as it was before. Apple surely didn't sell enough Macintoshs "to us", yet they survived. They sold them to other people that wanted to pay more. If you want to survive, don't create a game machine for people that don't have the money.