Digital Equipment Corporation was perhaps the second most important computer company in history, behind IBM. Its minicomputers challenged IBM, and, indeed, Unix first ran on a DEC PDP-7. DEC’s Alpha CPU was one of the few chips to make Intel nervous for its x86 line. It created the first really good Internet search engine. In a just and perfect world, DEC would still be dominating. Instead, it faded away in the 1990s. What happened to Digital Equipment Corporation, or DEC?
There’s a short answer and a long answer.
DEC in summaryDigital Equipment Corporation is a legendary company. It spend 25 years in the Fortune 500, peaking at number 27 in 1990 and 1993. Its peak revenue was $14.6 billion in 1996, but from 1991 to 1996 it lost money every year but one. In 1996, its peak revenue year, it lost $112 million. By 1998, DEC was #116 on the Fortune 500 list.
DEC made PCs, minicomputers, Unix servers, workstations, CPUs, networking chips, and the first great Internet search engine.
The short answer to what happened to DEC: CompaqThe short answer is that Compaq bought DEC in 1998 for $9.6 billion, a significant discount from its profits of that year. DEC had a lot of technology that Compaq didn’t really want, but Compaq bought it for the services business. Compaq saw services as the key to driving its future growth.
Compaq didn’t want the minicomputer business, or the Unix business, or the Altavista search engine, and it certainly didn’t want the chipmaking business. An upstart from 1982 picking apart its business, keeping what it wanted and selling off the rest like a corporate raider was a sad, inglorious end to a once-proud company.
The longer versionIn 1988, DEC was on top of the world. It was the second largest company in the industry, an investor darling, and everything looked great from the outside. But the company already had problems.
DEC’s biggest problem was responding too late to the rest of the industry. The original IBM PC was no threat to the DEC VAX, but 1986’s Compaq Deskpro 386 was. DEC also ceded too much of the workstation market to companies like Sun Microsystems.
DEC’s responseDEC’s competition came in fits and starts. DEC started a RISC project called PRISM in 1985 to replace the VAX architecture. It cancelled the project in 1988, and its lead developer, Dave Cutler, and much of the PRISM team went to Microsoft, where they built Windows NT. Soon after, DEC started work on the Alpha CPU. Windows NT ended up running on the Alpha, but DEC could have had control of it, rather than sharing it with Microsoft. In the meantime, DEC built a MIPS-based workstation that became popular. DEC didn’t transition that market smoothly to Alpha, effectively killing a successful product.
DEC mishandled Alpha too. Apple was interested in using the chip in what became the Power Macintosh line, but DEC never really took the offer seriously, creating a missed opportunity. Commodore wanted to put the Alpha in its ill-fated Amiga too. DEC was slow to line up second sources, and couldn’t produce chips as quickly or inexpensively as Intel. DEC had a successful chip business, between its Strong ARM CPU, Alpha CPU, and its Tulip network chip, which for a time was the best network chip for Linux machines. The Tulip was one of the first, if not the first, good quality inexpensive Ethernet chip and it ended up on mass-market cards from companies like D-Link, Linksys and Netgear.
DEC’s chip business would have been more successful if DEC could have simply produced more chips.
The downsizingThen DEC started downsizing in 1990. Over the course of six years, it cut its staff in half, at a cost of $4.8 billion. The loss of skilled workers cut its production capacity and cut the company’s reputation for stability.
DEC pushed out Ken Olsen, its controversial CEO, in 1992. Olsen didn’t get along with his board of directors and the board of directors didn’t understand the company business.
Altavista was the first commercially successful search engine, and Larry Page and Sergey Brin approached DEC in 1997 with their Pagerank system before founding Google. Page and Brin would have preferred to join Altavista.
DEC also invented the firewall. Computer security is a big business of course. It’s why I have a job, personally. DEC could have been in that booming field from the beginning.
Finally, there’s printing. DEC’s printing business was never as big as that of HP or Lexmark, but it was profitable. Whether DEC stayed in that business or sold it, printing was another missed opportunity for Digital.
Misconceptions about DECMany analysts say DEC was too slow to get into PCs. This is a bit of a misconception. DEC was right there in 1982 with its innovative Rainbow, which could do triple duty as a VT terminal, a CP/M machine, and an MS-DOS machine. Its problem was that it wasn’t completely IBM compatible. But DEC wasn’t the only one to make that mistake.
In 1989, DEC introduced a line of Intel-based PCs within its DECstation line and DEC got more and more aggressive with Intel-based PCs as the 1990s came on. Analysts frequently accuse DEC of positioning proprietary machines against PCs, but that was only part of its strategy. IBM didn’t compete against itself enough and maybe DEC competed against itself too much, but lack of a PC strategy wasn’t enough to sink the company. If DEC were around today, it’s entirely possible it wouldn’t be in the PC business anymore anyway due to the slim margins.
It’s also popular to criticize Digital’s reliance on VMS and Unix. As a security professional, I know how hard it is to get DEC VMS and Digital Unix out of environments. You can’t, even though 99% of the company wants it gone. Compaq didn’t want them after it bought DEC, and HP didn’t want them either after it bought Compaq, but both companies found them profitable. Had DEC, who did want them, continued, those businesses would have flourished in their niche. VMS would be even harder to get rid of if the company selling it actively marketed it and supported the two products’ cult following.
Ken Olsen: Punching bag
Ken Olsen’s overall message in 1977 was that the industry had work to do. And even as the home computer craze of the early 1980s rushed in, that work wasn’t done yet. Ken Olsen wasn’t the idiot people like to paint him to be.
The biggest problem with Ken Olsen was his loss of his mentor, Georges Doriot, in 1987. Without Doriot to advise him, Olsen couldn’t manage his relationship with his board of directors. And when the board of directors forced Olsen into retirement, they replaced him with Robert Palmer. Ken Olsen may have been the wrong person to turn DEC around in 1992. Robert Palmer, as it turned out, wasn’t the right person either.
It’s important to remember DEC wasn’t the only company in trouble in 1992. Wang Laboratories went bankrupt that year. IBM replaced its CEO in 1993 with Lou Gerstner. Even companies who didn’t replace their CEOs in 1992 like Compaq and Apple downsized significantly that year. Robert Palmer had six years to turn the company around and didn’t succeed.
What about Digital Research?Digital Research (DRI), maker of CP/M and GEM and active from 1976 to 1991, was a different company. Despite the similar names, the two companies were not directly affiliated. Digital Research borrowed the name and syntax of CP/M’s curious PIP command from a DEC OS, and DEC occasionally shipped products bundled with one or more of Digital Research’s software tools bundled like the DEC Rainbow, but that was it for the companies’ relationship. DRI sold out to Novell in 1991, and is its own story of what might have been.
What might have been with DECIt’s not hard to imagine an alternative universe where the Alpha ended up powering the Mac and Amiga. Whether an Alpha-powered Amiga would have been successful is another argument, but we know an Alpha-powered Mac would have been commercially successful. DEC didn’t know how to sell Alpha chips, but Apple would have known.
Meanwhile, Altavista could have easily become Google. That’s potentially bigger than all the rest of this combined.
And besides all of that, DEC had a successful and profitable services business, its own line of PCs, its minicomputer and Unix businesses, and should have had a booming Internet security business.
If DEC had realized what it had, it would have a good problem today: Being too dominant and having antitrust concerns. If DEC had restructured with better focus sometime in the early 1990s, it would have made fewer mistakes and at least preserved more of its shareholder value. A DEC with missteps likely still would have been involved in some mergers in the 1990s and 2000s. But it could have been a buyer instead of a seller.
IBM’s exampleIBM had to engineer a turnaround in the early 1990s, restructuring some businesses and ultimately leaving some businesses, such as PCs. There’s every reason to think DEC could have done the same thing. DEC wasn’t as big as IBM. But DEC had potential IBM didn’t.
IBM is no longer the company it was in 1998 when DEC sold out. Today in 2017, IBM is #32 on the Fortune 500 list. It was #6 in 1998 and it’s been sliding for two decades, but it’s still around. IBM no longer competes in Intel-based PCs and servers, the market DEC was criticized for being too little, too late in. Today, IBM survives on services, hosting, minicomputers, Unix servers, and mainframes. DEC had successful competing offerings in all those spots except mainframes and hosting. The hosting business didn’t really exist in the 1990s.
So the IBM example provides something of a worst-case scenario for DEC if it had survived: a mini IBM. Had DEC found success in one thing IBM couldn’t replicate, it could have been IBM’s equal. If not, it might have turned into HP. HP today isn’t too different of a company than DEC would have been, and it’s #61 on the 2017 Fortune 500 list.