On April 29, 1994, Commodore announced it was bankrupt and was going out of business. Its demise was a long time coming. Arguably it had been inevitable for 10 years. But the reasons Commodore went bankrupt are often oversimplified and misunderstood.
Loss of Jack Tramiel
Jack Tramiel left Commodore in 1984 as a result of a disagreement with its chairman and majority owner, Irving Gould. He resurfaced a few months later at Atari. Some argue Commodore’s demise became inevitable when Tramiel left. And there’s something to this, given Tramiel’s successors like Max Toy and Thomas Rattigan generally lasted less than two years.
But I’ll also argue Tramiel’s tactics of undercutting everyone in price weren’t sustainable. The Commodore 64 dominated the 1980s home computer market by undercutting everyone’s price aggressively, and the Atari ST had a nice run in the mid 1980s doing the same. But eventually someone else will figure out how to undercut your price.
It’s hard to imagine Commodore doing any worse with Tramiel at the helm for 10 more years. After all, Atari under Tramiel outlasted Commodore. But Tramiel alone wasn’t the answer.
Commodore went bankrupt because of the PC’s momentum
Others have said the PC’s momentum was too much for Commodore to overcome. But this too is an oversimplification. The Commodore 64 sold well for the entirety of the 1980s and the Amiga’s trajectory closely resembled that of Apple’s Macintosh.
The Amiga wasn’t going to overtake the PC but had every possibility of surviving as a niche platform. We have three incompatible standards for game consoles, so there’s no reason a third standard for computers couldn’t have also survived. The Amiga was so much better for analog video editing than any other platform that it could have owned that market even after the switch to digital.
Not keeping pace
The conventional argument is that the Amiga hit the market in 1985 with four-channel audio and 4,096-color video but was outmoded by 1989, when you could get a PC with 262,144-color video and 11-channel audio. Commodore didn’t respond with AGA until 1992.
This is all true, but it ignores the Amiga’s sales figures. The Amiga’s sales increased year over year until 1993, and the only reason it slowed down in 1993 was that Commodore didn’t have the money to produce more units.
Had Commodore been faster to respond with AGA or something better, no doubt the Amiga would have fared better. But the conventional narrative that the availability of VGA graphics and Ad Lib sound killed the Amiga overnight is an exaggeration. Commodore sold all the Amiga 1200s it could make, but it couldn’t make nearly enough of them to keep up with demand.
Commodore’s fab
I see this brought up on Reddit from time to time and frequently shot down, but there’s something to Commodore’s fabs. Commodore owning its own fabrication plants was an advantage until it wasn’t. It allowed Commodore to undercut everyone’s prices in 1981 and 1982, but Moore’s Law requires reinvestment to take effect. Commodore closed its fab in California rather than modernizing it, and last updated its fab in Pennsylvania in 1985. Even the 1985 update didn’t bring it to state of the art. Commodore licensed the rights to manufacture the 8088 CPU but never exercised it, because Commodore could buy 8088s for less than it cost for them to make them.
The antiquated fab meant the custom chips in the C-64 and Amiga couldn’t decrease in price as quickly as everyone else’s. Commodore switched to using contract fabs for the Amiga’s AGA chips, which was the right thing to do, but not getting its orders in soon enough meant they didn’t have enough AGA chips to meet demand for the Amiga 1200 and 4000 in 1992.
If Commodore had modernized its fabs, switched to contract fabbing sooner, or simply not botched the transition, 1992 and 1993 would have gone much better for Commodore. Any of those answers is a better answer than what Commodore ended up doing.
The fab problems were a long time coming
Commodore could see the problem with the fab a long way off. Commodore couldn’t keep up with demand for the C-64 in 1983 and 1984 because it didn’t have enough chips. Modernizing the fabs would have meant more chips per wafer, which would have meant lower prices and increased capacity. But what they did was close the California fab to save money in 1985. This ensured Commodore would never have the capacity to produce 3 million machines a year again.
Since Commodore could get by without the California fab in 1985, they should have modernized it instead, then shifted all production there and modernized the Pennsylvania fab at a later date.
Pricing
There’s an argument that Commodore priced the Amiga too high. Part of this was because their profit margins didn’t let them lower prices as aggressively as the clone makers. But I also don’t think Commodore should have gotten into a price war with the likes of Packard Bell. Given the Amiga’s dominance in the video production field, Commodore could have priced its big-box machines at the high end to increase its overall margins, much like Apple charges superpremium prices for its desktop models. Newtek actually raised the price of its Video Toaster over time, and its customers simply paid the new price. Newtek realized they were competing with machines that cost several times what the Amiga did, and took advantage.
The Amiga 500 and 2000 grew less cost competitive over time as they aged, especially if you wanted a hard drive. But the Amiga 1200 and 4000 closed the gap by using cheap commodity IDE drives like everyone else.
Irving Gould
We really can’t heap too much blame for Commodore going bankrupt on Irving Gould and his yes-man, Mehdi Ali. Gould was an intelligent financier, but technology wasn’t his thing and he wasn’t interested in learning. For him, Commodore was a piggybank to loot for however long it lasted, while he was in his shipping and vending machines businesses for the long haul.
Gould paid himself lavish salaries while Commodore struggled, maintained redundant office buildings in three countries and traveled between those three countries at Commodore’s expense, all so he could evade taxes. All of this benefited Gould tremendously, while draining cash that Commodore could have been putting to profitable use. Such as building Amiga 1200s.
Commodore ran out of time before Gould did. Commodore died April 29, 1994. Gould died December 17, 2001, aged 82. At the time of his death, he was still fighting off lawsuits related to his mismanagement of Commodore.
David Farquhar is a computer security professional, entrepreneur, and author. He has written professionally about computers since 1991, so he was writing about retro computers when they were still new. He has been working in IT professionally since 1994 and has specialized in vulnerability management since 2013. He holds Security+ and CISSP certifications. Today he blogs five times a week, mostly about retro computers and retro gaming covering the time period from 1975 to 2000.
It’s absolutely fair to lay the blame for Commodore’s demise on Irving Gould. Although it’s interesting to note Commodore would likely not have survived long enough to acquire MOS and enter the computer business had it not been for Gould’s intervention in the late 1960s. In researching one of my own articles on Commodore history, I reviewed the lengthy report prepared by the Canadian government in response to the Atlantic Acceptance scandal. As relevant here, Jack Tramiel made an ill-advised decision to purchase a retail store chain–a mistake he later repeated at Atari Corporation–and was then left with a $3 million short-term note he couldn’t pay once Atlantic Acceptance, his main source of cash, collapsed.
It was Gould who bailed Tramiel out by finding a buyer for the retail chain. Absent intervention from Gould or someone else, it’s likely Commodore Business Machines (Canada) Limited would have followed Atlantic Acceptance into receivership itself. The Gould-led Commodore International would never have existed.
Yes, the irony of Gould building Commodore up and tearing it down and having no idea what he could have had is significant.
when you could get a PC with 262,144-color video and 11-channel audio
sound blaster was 11-channel audio?
did Commodore switch from 68k to Power PC or ARM?
could Commodore licensed clones and stay in business
or just use off the shelf parts and sell Amiga OS
No, Commodore didn’t survive long enough to move to another CPU. Selling clones would not have helped, the same way it didn’t help Apple. And switching to off the shelf parts would have been a very long transition, the thing that made the Amiga so special was its custom chips. And it was Ad Lib that was 11 channels.
well okay thanks.
so why did Apple Mac survive even though it was originally monochrome had a 9 inch monitor, was it because its GUI was superior to Atari/Amiga?
I recall that IBM wanted to offer a common powerpc platform that any cloner could copy.
did Motorola ever offer a cloneable 68k reference platform, no royalties?
how come ARM Acorn Archimedes wasn’t sold here in the USA and why didn’t Apple or Amiga/Atari consider ARM instead of 68k
in my childhood fantasy Tandy Radioshack, instead of selling the Tandy 1000 or Color Computer, offered a 68k or ARM computer with just off the shelf parts and cloneable. it made money selling deskmate and OS/9 and GEM
Why Apple survived is complicated, very much off topic for Commodore and Amiga, and I’ve written a lot of blog posts about that. Short answer is they were the last to run out of money, and soon after Microsoft bailed them out in the 90s, Steve Jobs figured out he needed to shift to an infinite game.
IBM and Motorola did offer something called the CHRP. A few clones using it did reach the market, generally running Windows NT, but they weren’t popular.
i with tandy sold a 68k for $799 1mb in 1987
why was Apple mac se so much more expensive than a Commodore and Amiga or Atari st but similar specs
> “… video editing than any other platform that it could have owned that market even after the switch to digital.”
Unfortunately, this isn’t correct. The Amiga’s success in video was almost entirely due to the Video Toaster. However, the way the Toaster achieved its extraordinary value vs traditional television production gear was by being *tightly* coupled both with the Amiga’s custom chips AND with NTSC analog composite video. The Toaster was a brilliant hack to create the miracles it delivered. The trade-off was it could never work on any other computer and it could never work on any kind of video that wasn’t exactly, precisely old-school NTSC analog composite video. It relied on the arcane pixel clocks, front porch timing, interlace and color sub-carrier of broadcast television standardized in the 1950s in the RS-170A standard.
This is why, despite its incredible success, the Toaster was never ported to another computer, there was never a component video version, and there was never a PAL Toaster for Europe. It wasn’t due to lack of demand or desire. It was literally impossible. The Toaster was so innately analog, NTSC and Amiga all the way down to its DNA, that changing literally anything about those complex analog timing relationships would break ALL of the Toaster’s magical abilities.
I can speak authoritatively on this since I worked at NewTek during the entire creation of the Toaster and for years after it shipped. Also, to a lesser but still significant extent, much of the Amiga’s graphical prowess was tightly coupled to 15khz standard definition video. Even just de-interlacing the Amiga to 31Khz involved some non-optimal downstream processing that created its own issues. That’s why Commodore was working on retargetable graphics toward the end. The brilliant custom graphics chips that so defined and differentiated the Amiga in 1985 were integral to its DNA but became barriers to transitioning to the future by the mid-90s.
Sadly, in a digital video world all the analog video production magic of the Toaster and the Amiga (thanks to its analog video synced clock speed (7.16 Mhz)) just evaporated. In the digital video realm both NewTek and Commodore would have to completely start over with no prior advantages other than brand. NewTek was able to ultimately transition to the PC with the Tricaster but it required several years and creating a new technology platform completely unrelated to the analog Toaster.
I bought an Amiga 500 back in 1990 – I was running a Software Etc. in Eden Prairie, Minnesota. I’d started as a part-timer in their Edina store, which was also near their headquarters. Edina was store #5, very high volume, usually only behind the 5th Avenue store in New York. When I got transferred to EP, that store was already 30% below sales plan for the fiscal year which started in October.
I used a trick I learned in Edina-The store manager there had a small budget to order Mac-specific software which sold. I decided to use our inventory in our warehouse to specialize in Amiga software. I would request any Amiga titles, kept in-stock on other stuff like top-selling Windows products, and we ended our fiscal year 33% over our total year sales goal.
While they did bring in various video game stuff that led to the chain morphing into Game Stop, we had the highest percentage of sales of Amiga software and accessories, which helped me hit that goal. So I owe Commodore a lot for that year’s bonus…
I think it took Commodore a while to go bankrupt, considering the many mistakes it made along the way. The most serious of these was its refusal to read the market and adapt in time; when it realized this, it was too late.
It was unacceptable in 1984 for them to create new computers with closed architectures when the PC was growing at an alarming rate in all world markets. Even worse, each new computer line was completely incompatible with the previous one.
In 1984, it would have been insane for anyone to buy the Commodore Plus/4 instead of modern PCs or the Commodore 64, which already had millions of applications.
The demise of Commodore is one of the most fascinating topics in the history of the computer business. And as an Amiga owner at the height of that machine’s UK popularity in the late 80s/early 90s, I lived through it (OK, I was just a kid, but still…)
A few straws in the wind which haven’t been raised in the article – Commodore lost a ton of money on a succession of disastrous product launches at the start of the 90s. The CDTV we can excuse to some extent, because Philips made a similar bad bet on the CDi. But the Amiga 600 was poor in both concept and execution, and destroyed sales of the still-successful A500.
At root though, Commodore’s problem was that it wasn’t a computer company. It was a widget maker. Commodore sold typewriters, calculators and thermostats before computers. The company’s corporate leaders had no vision, ever, beyond selling stuff for a short-term profit. It never had a strategy, a business plan, a product roadmap, a mission statement. And because of this they never really capitalised on the success of the C64 or the capabilities of the Amiga.