According to 'Commodore: The Final Years' (whole trilogy highly recommended, BTW), page 129, 'Jeff Porter realized it would not be possible to significantly cost reduce the Amiga 500 to get it into the $250 retail price range'.

This is really important. All the flailing around Commodore did in their final years, designing a bunch of different machines that missed the point. I used to wonder why it didn't occur to them that they needed to focus on cost-reducing the A500, get back to selling a good computer at a surprisingly low price, that was always their core strength. Well according to the above, it did occur to them, but they decided they would not be able to do it. And everything else the company did in the last few years of its life, was the rebound from that. (Okay, I exaggerate slightly. Some of their higher-end products would have been important anyway. But nothing else had the potential to sell millions of computers.)

Why was such cost-reduction not possible? Two possibilities come to mind:

  1. The late eighties was the time of the RAM famine; half a megabyte was significantly costly all by itself. (But then why did the verdict not change when the famine ended?)

  2. There were a lot of chips on the board. They needed to be fused into fewer chips. But Commodore Semiconductor Group (nee MOS Technology) had been starved of capital to keep up in process technology, and could not make the larger chips that would be needed.

Is there something else I am missing?

  • 1
    You should take to heart the answer most people are down-voting below...Commodore failed because they couldn't make sufficient profits from each Amiga, not because Amigas were too expensive. Same downfall as any company that thinks they can "make it up on volume".
    – Brian H
    Commented Jan 19, 2020 at 16:34
  • @BrianH I didn't downvote that answer, but I don't think it negates my point. Commodore was indeed not making enough profit from each Amiga - which was because they didn't have the marketing position to sell for a higher price, but equally because each one cost them too much to make. Profit is revenue minus cost.
    – rwallace
    Commented Jan 19, 2020 at 16:47
  • 1
    I get your point that higher margins need to be carved out from expenses too. I just think that Commodore product engineering and consumer value were always quite good, but the company management was quite poor.
    – Brian H
    Commented Jan 19, 2020 at 18:06
  • @BrianH I certainly agree with that!
    – rwallace
    Commented Jan 20, 2020 at 23:16

3 Answers 3


'Jeff Porter realized it would not be possible to significantly cost reduce the Amiga 500 to get it into the $250 retail price range'.

They could have cost-reduced the A500 - perhaps even to $250 retail - but they would have had to make some compromises that (thankfully) they weren't willing to do.

There were a lot of chips on the board. They needed to be fused into fewer chips.

The custom chipset was only 4 chips (Agnus/Denise/Paula/Gary). You could keep those chips and still make significant cost savings in other areas. First step would be to remove little used features such as the serial and parallel ports, CPU expansion bus, and internal RAM expansion board / realtime clock connector. This also eliminates a lot of passive components that were required to meet EMI specs etc., and significantly reduces the board size and cost.

Without serial and parallel ports the essential I/O functions can be combined into a single CIA chip. Then all the custom chips can be made surface mount and soldered onto the motherboard to get rid of troublesome IC sockets. Finally the modulator can be integrated onto the motherboard, or you could simply provide a basic RGB port and make users buy a separate monitor (not included in the base price!) or use a SCART TV.

But for really big savings the answer would be to leverage the PC clone market. With small tweaks your cost-reduced A500 could use a standard AT keyboard, floppy drive, case and power supply - supplied by whichever Taiwanese manufacturer was cheapest - or you could just sell bare motherboards and let the purchaser choose what they require to complete it! (perhaps nothing if they can use bits from an existing PC).

Well according to the above, it did occur to them, but they decided they would not be able to do it. And everything else the company did in the last few years of its life, was the rebound from that.

Yes, it seems that they considered, but rejected, going down that path. The A600 is commonly accused of being a failed attempt at producing a cost-reduced A500, but this may have had more to do with manufacturing costs in general rather than the design itself. The inclusion of IDE and PCMCIA ports shows that they wanted to do more - and those extra features justified a slightly higher price. If the A600 had been truly cut down to the bone it might have been cheaper, but much less desirable.

The closest Commodore got to a true cut-down Amiga was the CD32. No keyboard, serial, parallel, RGB ports etc., essential I/O functions integrated into a single custom chip - yet they still (thankfully) provided a keyboard port and expansion bus. So even the CD32 wasn't as cut down as it could have been. Clearly even at this late stage Commodore held out hopes of the Amiga being more than just a games machine shoved out at the lowest possible price.

  • 1
    Or maybe instead of a full motherboard, the Amiga could have been some kind of accelerator in an ISA or PCI slot, with its own video out. Commented Jan 19, 2020 at 17:42
  • 2
    By removing serial and parallel ports you lose the ability to connect printer and modem. That really would have been a games machine shoved out at the lowest possible price! And what did the expansion bus cost? It was just a piece of PCB, they even pushed the cost of the plug onto the peripheral manufacturer.
    – user6580
    Commented Jan 19, 2020 at 22:59
  • 2
    @AdamEberbach Even a gold-plated edge connector costs money. Also more traces so the PCB might have to be larger and/or have more layers (=$$$), a custom case (if used) would need a hatch, and extra filtering might be required to reduce EMI (the A1200 had several bodges applied to reduce EMI, which caused problems with some accelerator cards). Finally, with expansion devices not supported the 'glue' logic could be simpler, development and testing costs reduced, and reliability improved. Commented Jan 19, 2020 at 23:30
  • 1
    @AdamEberbach the CD³² still supported serial data transfer. The data lines were connected to the keyboard port. With an appropriate adapter, you could use it and still use the keyboard. So the costs for a serial port were shifted to the few percent of CD³² owners who (like me) actually wanted to use a serial connection. If USB only was a bit earlier…
    – Holger
    Commented Feb 21, 2020 at 9:39

They did, with the A600. But in true late-stage Commodore fashion, they screwed it up and made it more expensive.

To cost-reduce an A500, you'd have to reproduce its spec on simpler silicon. The market wasn't interested in an 8 MHz 68000 in 1991/92: the PC had stolen all of the Amiga's thunder at commodity prices. The Amiga's niche silicon was just too expensive to produce and couldn't benefit from commodity use from multiple vendors.


Commodore obtained an infusion of cash from Gould, which Tramiel used beginning in 1976 to purchase several second-source chip suppliers, including MOS Technology, Inc., in order to assure his supply.[6] He agreed to buy MOS, which was having troubles of its own, only on the condition that its chip designer Chuck Peddle join Commodore directly as head of engineering.

They bought a company to make their own chips, "which was having troubles of its own", and on top of that they stole the guy in charge. Strikes, one and two.

In 1983, Tramiel decided to focus on market share and cut the price of the VIC-20 and C64 dramatically, starting what would be called the "home computer war". TI responded by cutting prices on its TI-99/4A, which had been introduced in 1981. Soon there was an all-out price war involving Commodore, TI, Atari, and practically every vendor other than Apple Computer.

Why was that? Because Apple doesn't make anything. They take off-the-shelf parts, purchased from the lowest bidder using buying power alone, and have things built for them in countries with lax labor laws. Strike three, you're out.

  • 9
    Citation needed on Apple’s manufacturing process circa 1990. Also it’d be more accurate to say that Apple used simple custom chips only — the video and audio parts were custom, the drive controller was custom, they were just immensely more simple than the Amiga parts.
    – Tommy
    Commented Jan 18, 2020 at 21:59
  • 8
    Your use of the present tense in the last paragraph is correct. In 1980-85 manufacturing of circuit boards was in Singapore and everything else was made in Carrollton, TX, Fremont, CA, and Cork, Ireland, as well as several smaller sites scattered around the US and Ireland. In fact, the period in question was when Apple was still trying (and failing) to make US manufacturing of personal computers a good idea... Commented Jan 18, 2020 at 22:01
  • @Tommy - that's a tough one. The 90s were a bad decade for Apple, and when they switched from Motorola to Intel, it was a bad day for Motorola. If you buy so much stuff from Motorola, (in their hey days of the 90s), that you could hurt them financially, I think it's ok to say they don't make their own chips. 'Custom' perhaps, but not from a factory that says Apple on the wall, afaik.
    – Mazura
    Commented Jan 18, 2020 at 22:23
  • 4
    The chips that Apple bought from Motorola were exactly the same as the chips that Commodore bought from Motorola. I think that if you’re somehow desperate to pull Apple into it, the difference is that the Macintosh was never defined in terms of hardware. Absolutely any machine with a 680x0 can be a Macintosh with the proper ROM. No need to figure out how to cost reduce the chipset, just throw together a new one and let the software guys fix it.
    – Tommy
    Commented Jan 18, 2020 at 23:40
  • 1
    To extend @Tommy 's point there were boards for the Amiga and Atari ST like the Spectre GCR to use Macintosh ROMs to run Macintosh software. Given the relatively high level of classic Macintosh software most programs could work on any 68k machine. Commented Jan 24, 2020 at 15:49

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .