Once-Mighty SGI Sold to Rackable for $25 Million

Paul Sakuma/AP

A collective shudder rippled through Silicon Valley on Wednesday morning, as Rackable Systems announced its purchase of Silicon Graphics Inc. for just $25 million in cash.

If you travel all the way back to 1997, SGI was pulling in close to $4 billion in revenue a year. The company produced some of the flashiest computers on the planet for handling tough graphics jobs. SGI was the next big thing in Silicon Valley, and people adored the company.

For some perspective, consider this sentence from a 1998 story in The Times.

Although Silicon Graphics is not considered a technology bellwether like Intel or Compaq, it was once one of the nation’s fastest-growing companies, best known for building extraordinary computers that helped create special effects for the “Jurassic Park” films.

Ah, those were the days.

Terms like “bellwether” and “fastest-growing” haven’t been used anywhere near the SGI name for years. The company used to thrive by selling computers based on its own chips and operating system. Such technology, however, was undercut by cheaper graphics products from companies like Nvidia and cheaper mainstream chips from Intel.

In addition, SGI made a blunder by opting to move all of its computers over to Intel’s lackluster Itanium chip.

SGI was forced into bankruptcy a couple of years back and has been struggling ever since. (The company announced that it had filed for bankruptcy once again on Wednesday morning just before revealing the Rackable deal.)

Rackable sort of resembles SGI, in that it used to be the next big thing in corporate hardware as well. As I reported earlier this week, Rackable was a shining star in 2006 before Dell started putting serious pricing pressure on the upstart, fighting it for customers like Facebook, Microsoft and Yahoo.

By buying SGI, Rackable takes on some engineers with expertise in building large, complex systems as well as some intellectual property around graphics and server technology. (Regrettably for SGI, the company sold off some its key 3-D graphics technology to Microsoft earlier this decade.)

While SGI has met with an unceremonious end, one of its founders has a new beginning.

The company was started in 1982 by former Stanford University professor James Clark, who went on to found Netscape Communications (since we’re going down memory lane). Last month, Mr. Clark, 65, married the Australian swimsuit model Kristy Hinze, 29.

Comments are no longer being accepted.

Nerd marries a model half his age….so the story does have a happy ending!

Coleman McCormick April 1, 2009 · 9:59 am

I remember my dad’s engineering firm using exclusively SGI machines back in the early/mid-nineties. They used to be so one-of-a-kind and impressive.

Then they disappeared around the time of the tech bubble.

This is a 363 sale. If another bidder comes along at a higher price in bankruptcy court — entirely possible at this value — Rackable will not get the company.

Microsoft bought SoftImage which was the major competition to SGI’s Alias/Wavefront Animator, then ported SoftImage to Windows and destroyed a lot of SGI’s market. SGI not only moved to Intel, but offered boxes running Windows instead of moving to Linux.

It was a huge blunder for SGI and gutted their market. They did eventually make the move to Linux but it was too late. A few years later MS sold off SoftImage.

Amazing end to SGI … we need to find that visualization software and re-release it!!!!!!

I used to work with SGI computers and I can tell you the reason they went out of bussines is because they were unable to adapt to the needs of their costumers. Their computers were ubelievably expensive, few companies were able to add them to their annual budget and the depreciation ration was over 50%. The software was incredibly cumbersome and absolutey impossible to master because they kept changing it. And they were snobs. As soon as Electric Image and other companies came up with Mac and Windows versions of 3D programs SGI lost its edge. Pretty much the same story with Netscape…bad management and poor leadership.

I am sitting here looking at my original Apple Newton which has an SGI visitor sticker on the back from 1995.

My visit to SGI to discuss creating virtual reality walkthroughs for stadiums and real estate were exciting yet led nowhere.

SGI was *way* cooler than Apple to me at the time. I had an O2 machine which was pretty powerful.

I remember when all of Hotwired (WIRED magazine) ran on one SGI Indy.

Seems like a series of bad decisions and missed oportunities tanked the company.

Difficult to think that the same thing could happen to nvidia in a few years.

It is a sad story about SGI. But I find the last paragraph about Jim Clark the most interesting.
There is always hope. If you can make a lot of money using your brains, you may be able to take home a beautiful woman!

No mention of Apple hurting their market? C’mon, a modern Mac tower with it’s ease of use and good customer service, along with their marriage to Adobe and others, brought the arrogance to it’s knees. Reminds me of the death of Scitex and Quark.

Hire good customer service people, and be light on your feet. The world is always changing.

I have a great deal of experience with SGI, spanning fifteen years. I was shocked when the Board put Bo Ewald in the leadership spot a few years ago. He has no real name recognition, and a reputation for sinking companies.

I sincerely hope that Eng Lim Goh, Bob Pette, et al are partnered with a solid company that realizes the value of visualization…and allows the SGI originals to get back to building state of the art viz products.

Agree with others that the failures of the past few years have been down to poor leadership (IMO, at the senior executive level). And let’s not forget, someone with deeper pockets and a greater vision may come along and make this company great again.

Many of us hope so.

SGI was the best company I ever worked for. I have never been surrounded with as much creativity and brain power.

Sorry to see them go by the wayside.

“If you travel all the way back to 1997″

Your date is off and in 1990 SGI was already the leader in computer graphics. I still remember seeing in 1990 one of their programs with a cube of jello being deformed as it ricochets off walls as it is moved by the user.

Can only think that SGI got sidetracked by trying to enter the general purpose market and not sticking with it’s lead with those who needed top of the line graphics.

Juan Pablo Reyes Altamirano April 1, 2009 · 4:32 pm

I was only a kid when I first heard the words “Silicon Graphics” uttered in the making of Terminator 2. Eighteen years later, with three O2’s, one Octane and experience in working on an Onyx 350, I feel as though a very dear friend has just passed away.

OpenGL, Nintendo 64, Final Fantasy VII, Power Animator, Maya, OpenGL Performer…and a gazillion other things, including the internet as we know it, owe a huge debt of gratitude to these visual computing visioneers (Well, Jim Clark got away with happiness so our debt with him is pretty much covered).

SGI, DEC, CDC…this list of bygone acronyms is getting longer.

I used to work with SGI computers and I can tell you the reason they went out of bussines is because they were unable to adapt to the needs of their costumers. Their computers were ubelievably expensive, few companies were able to add them to their annual budget and the depreciation ration was over 50%. The software was incredibly cumbersome and absolutey impossible to master because they kept changing it. And they were snobs. As soon as Electric Image and other companies came up with Mac and Windows versions of 3D programs SGI lost its edge. Pretty much the same story with Netscape…bad management and poor leadership.
— S. Gee Ian
******************************************
I do not about being snobs, but back in 1990 the description of SGI computers could be probably true for all computer manufacturers.

The difficult part for these companies was that they did not understand that creating hardware is an easy problem to solve but finding the software experts to use the hardware is a difficult problem.

In reality we were better off in the difficult problem of a lack of software experts in 1990 than we are today. The current idea that you can deal with the problem of not enough software experts by using cheap labor makes as much sense as using cheap labor to solve the problem of not enough brain surgeons. This solution has only decreased the supply of software experts as only fools would want to enter a field such as software development if the work is considered that of cheap labor.

and yet my SGI Origin 350 continues to chug away, having only rebooted twice in 4 years, once due a lightning strike on my building that destroyed the UPS, the other time when we moved to a new building.

A company way ahead of their time, sorry to see them become just another name for the IT Graveyard.

The thing I remember most about SGI (as an outsider) was their involvement in the burgeoning Hollywood special effects market. Maybe it was mismanagement; but it seems to me that their business of specialized workstations was swamped by the advent of massively parallel commodity processors doing the same work, with the specialization taken over by the software.

Aahh it is a sad day indeed! Working at SGI was a truly magical experience, changed my life…such special friends. What a loss!

I remember the day the Lead LNXi Sales guy brought in “Bo” and introduced him to the Director as the man that was going to turn LNXi around. Sadly, I also remember the day the Lead SGI Sales guy brought in “Bo” as the man that was going to take SGI back to its roots and re-engineer it into the great company it once was. The only thing different in the room was the name of the Sales guy.) In the words of a truly great man; “This is like deja vu all over again.” But in all fairness Bo really did do what he said he was going to do. Remember when SGI was short for Silicon Graphics, Inc, and then someone re-branded it SGI. Under Bo’s leadership it was re-branded Silicon Graphics, Inc.

Morris

Brad, I think you’re confused about the timeline of SGI and Windows NT. This was years ago, well before Linux was a sophisticated multi-proc system. It was NT versus IRIX then, and IRIX has a reputation for mediocre stability.

I believe SGI’s mistake was in announcing they were going to support Windows NT and then changing their mind and dropping it. UNIX dogma instead of hard-nosed business prevailed, and this (according to our SGI sales rep at AT&T) went against big customer demand for NT.

Ah, SGI — I used to sneak into their booths at SIGGRAPH to see what the next hot 3D machine would look like. But let us not forget Evans & Sutherland, true founders of 3D graphics, who again, like so many outstanding engineering companies, lost their way through mismanagement and a lack of insight into market changes and customer demands.

I was employee #764 and started with SGI in 1984. There was no other place in the industry that was as exciting as working there at the time. We were going to change the world and I think we did for quite a while. The people were what made it what it was, a creative environment, and the technology was what enabled the people.

Unfortunately the money earned by many of the creative folks was not enough. They started leaving in droves to move on to other ventures by the early 90’s. SGI was just riding on it’s own momentum by then. I was one of those pepople that left, following the ‘smart’ guys out the door. We started a company along with some DEC people named MASPAR. We failed as did most of those start ups at the time. Soon after the first colapse of SGI took place. The rest is history.

For the SGI old timers, and I still have the first edition lavender T-shirt with logo, we failed to crush SUN. That was one of our focuses back in the day. It looks as if SGI was finally eclipsed.

oh, yeah. remember when…

remember when we tried to tell them NOT to build intel-based workstations? they were too smart to listen. remember when we questioned the new logo and “sgi” font (shown in the photo at top)? remember when we suggested basing a server line on a non-existent intel processor line might be a problem? remember when we asked why a company would want to build workstations in Switzerland, and then we found out the our Chairman had finagled a free Swiss mansion? Geniuses at the helm…all the really smart guys went un-thanked — so they went to Nvidia.