Why did IBM fail at PCs?

If you ask why did IBM fail, I assume you mean why did IBM ultimately fail in the personal computer market. IBM is still in business, after all. But its exit from the PC market after 24 years, including a period of dominance in the 1980s, does seem curious. And it raises another question: What does IBM do now?

I experienced IBM’s fall in this market firsthand. I sold computers at retail in 1994 and 1995. IBM’s computers at that time were no worse than anyone else’s, but I had an extremely difficult time selling them. Many consumers didn’t trust IBM and didn’t want to get somehow locked in. There was nothing wrong with those machines, but it sure was a lot easier to just sell them a Compaq.

Now that I think about it, the best thing we probably could have done to sell IBM PCs in those days would have been to connect as much non-IBM stuff to them as we could.

Understanding IBM

To understand IBM, it helps to remember they always were more than a PC company. That said, they have evolved over the years.

The IBM of the 1980s was a computer hardware company. Whether you needed a desktop computer for your office or a mainframe big enough for every employee in a Fortune 500 company to use, IBM had you covered. They had all that and everything in between.

In the early 1990s, IBM made a painful transition from selling computers to selling services and software. Services and software proved to be more profitable. For a time, IBM kept selling PCs because it made it easier to sell services. But eventually that stopped being enough of an advantage, so IBM sold off its PC line in 2005 to Lenovo.

The rise of the IBM PC

The IBM PC, in its standard early 1980s configuration
IBM built its original IBM PC, introduced in 1981, using mostly off-the-shelf parts.

The rise of the IBM PC has been discussed at length, so there’s little need to go too far into it here. Suffice it to say, it legitimized the personal computer market and it was far more successful than even IBM expected.

But IBM got greedy. A clone market popped up around the IBM PC, and IBM didn’t want to share. That’s a story of its own. But essentially, IBM created an open standard like VHS videocassettes or Blu-Ray. Perhaps they left it a bit more open than they intended.

The chink in IBM’s armor

IBM wanted a successful PC business. But IBM didn’t want it to compete with its minicomputer business. So in 1985, when Intel released its 386 CPU, IBM had no intention of using it. To IBM’s way of thinking, PCs belonged on desks. If you wanted a server, you could live with a repurposed 286-based AT. Or you could buy one of IBM’s costlier minicomputers.

This left an opening in the line. Compaq walked right in, and so did several other smaller-name PC makers. These early 386s didn’t take full advantage of Intel’s new chip, but they were nicer than a 286. This forced IBM’s hand. So IBM responded, but there was a catch.

The IBM PS/2 – the beginning of the end

IBM PS/2 Model 30
This IBM PS/2 Model 30 featured an 8086 CPU and replaced the PC/XT in IBM’s product line. The fastest PS/2 from 1987 was the Model 80, which featured an Intel 386 CPU running at 16 MHz.

IBM replaced the popular PC line in 1987 with its PS/2 line. The Personal System/2 was a technical improvement over the PC, but it also moved the architecture from an open system to a very closed system. The main system bus was called Microchannel, and you had to pay a royalty to IBM to use it. That meant IBM made money whether you bought a PS/2 from them, or a PS/2-compatible system from someone else. It also meant IBM made money on the expansion cards.

Consumers saw this vendor lockin from IBM as overly aggressive and anti-consumer. The PS/2 line didn’t exactly flop, but it didn’t set the world on fire either. And rather than pay royalties to IBM, Compaq and eight other companies created an alternative standard called EISA that they kept open. So essentially the rest of the industry ended up cloning Compaq instead of IBM.

In the 486 era, EISA mostly gave way to VESA, another mostly open standard created by an industry consortium. When the time came to replace EISA, Intel took the lead on creating PCI, a standard still in use today.

IBM went into the PC market thinking they’d do really well if they could capture around half of it. By 1987, IBM was greedy and wanted nearly all of it. That backfired badly.


By the early 1990s, IBM realized they were missing out on this huge open market, so IBM started releasing computers that used the non-IBM standards instead of Microchannel, such as the PS/1. By the time the Pentium was on the market, most IBM computers were using open standards again with the exception of the last of the PS/2 line, like the Pentium-based Model 95.

IBM’s reformation and embrace of open standards helped, but by then they were just another PC maker. Microsoft and Intel drove the industry, so by 1995, you didn’t hear the words “IBM compatible” very often anymore. Instead you heard “Wintel,” for “Windows” and “Intel.” In 1992, IBM considered buying another PC maker, such as Northgate, Zeos, or Gateway 2000 to gain market share. Instead, IBM opted to launch a subsidiary it called Ambra, selling computers direct using Gateway 2000’s model. IBM’s goal was to capture a modest 10 percent of the direct-order market, but Ambra failed to do that after two years. IBM closed Ambra in 1994.

In the consumer space, people remembered IBM’s moves in the 1980s and some, at least, remained nervous. IBM’s reputation was Packard Bell-like, even if the quality wasn’t. The lone exception was the Thinkpad, a line of stylish, well-built laptops that were so good people were willing to put aside their fears about IBM for it.

IBM never built a desktop PC with the Thinkpad’s cachet. Then again, nobody else did either.

PCs becoming commodities

It’s entirely likely that IBM would have eventually gotten out of PCs anyway, even if the PS/2 had been a raging success. IBM used to make printers, memory chips, other computer chips including CPUs, and hard drives. Over the years, IBM spun off or sold all of those businesses. IBM’s philosophy has been that it’s better to get out of a business a year too early than a year too late.

And PCs are a hard business to be in. Dell had financial troubles and took itself private in part to avoid angry shareholders. HP split itself into two companies to try to deal with its lack of profitability in PC sales. Gateway and Emachines merged and then sold out to Acer.

The only scenario I can see working for IBM would have been if it had been able to retain premium brand status, like Apple, selling its PCs at Apple-like prices.

The margins worked for Lenovo. But Lenovo’s business model is different from IBM.

A segue into business theory

There is a classic business book called The Goal. Its core message is that businesses exist to make money, and they often lose site of that core goal. Businesses can lose their way by doing things for the sake of doing them.

In The Goal’s example, the company lost its way by diving too deep into robots, without thinking about whether those robots were helping them make money. The robots were cool, but they weren’t helping the bottom line.

Ultimately, IBM management looked at the PC business and came to the conclusion it was like the robots in The Goal. Being a big name in the PC market might have been nice, but it wasn’t making the company money. So they sold it to a company who thought they could make money at it.

What does IBM do now?

IBM is still very much in business, and they do still make computer hardware. It’s just a different kind of computer hardware. They make minicomputers, mainframes, and AIX servers. AIX is IBM’s own proprietary Unix. All of these are a much higher margin business and it allows them to sell software and service too. IBM is a very different company today than it was in 1981. Some people aren’t too bullish on its future, but it has proven adept at reinventing itself.

That’s why I’m uncomfortable saying IBM failed. IBM is a different business today than it was in 1981 or even 1995. But it’s still in business, selling things not everyone even realizes still exists. It’s news to many people that mainframes still exist.

And if you were to ask someone who works at IBM, “What does IBM do now?” they might not even say hardware. IBM does a lot of hosting and consulting now. If you want someone to run your datacenter or part of your IT department for you, IBM is one of the companies you call.

IBM laid off a lot of workers in 2016 and again in 2017. But what I expect to see going forward is IBM continuing to shed unprofitable business lines and continuing going forward, even if it means being a smaller company. I don’t see the company going completely out of business.

One thought on “Why did IBM fail at PCs?

Leave a Reply

%d bloggers like this: