Rackable Systems Eats Silicon Graphics

When I moved to San Francisco in the early 1990's and started meeting the computer graphics crowd, the place everyone said they wanted to work was Silicon Graphics. Not only was that where the cutting edge lay, it was also seen as a smart career move. Surely, in a few years, Silicon Graphics would surpass HP and then Intel and IBM.

It did not work out that way. Today Rackable Systems (RACK) announced it would buy Silicon Graphics (SGIC) for about $25 million. While that is certainly a bargain basement price for a treasure chest of intellectual property, I cannot assume that Rackable will do well with the acquisition.

Silicon Graphics' (aka SGI) traditional problem is running a business that actually makes money. That is exactly the problem Rackable Systems has had these last few years. Rackable has a substantial amount of cash in hand, but it was mostly generated from its initial IPO, not from profits. Rackable grew rapidly and had some profitable quarters a few years back, but then they got shoved around by bigger competitors, notably Sun, HP, Dell, and IBM. I think Rackable on its own could be profitable again if IT spending picks up some time this year, but this acquisition will create distractions and a lot of new variables to manage. So we have the dilemma of a company with a tremendous opportunity that could turn into a serious downside.

What are Rackable shareholders (that includes me) getting for their $25 million in cash? The latest results for SGI showed revenues of $82.8 million and a GAAP net loss of $49.2 million and an EBITDA of negative $17.3 million. That is no way to run a business. And it is not just from the recession. Once upon a time SGIC's market capitalization was over $1 billion.

So the first thing Rackable management should do is can as much of SGI's management as is necessary. Cut costs to the bone. Assuming an $80 million per quarter run rate, if you can't generate $8 million of profit on that, you should get out of the business. If there is revenue that can't be made to generate a profit, jetison it.

Having watched variously poorly managed technology companies for years, I will ball park what I would expect to see if I were a Rackable manager reviewing SGI. Sure, if you spend $1 million on a product, and it is a good product, and offer it for sale at $800,000, someone is going to buy it. What you need to do is price the product at $1.1 million. If your customer will not pay $1.1 million for the product that cost you $1 million, you don't need that customer. So first you try to raise the price, and if that does not work, you close that product line down. Let the managers and workers find jobs elsewhere (good luck). Or if they want to keep their jobs, and sell their product for $800,000, they need to find ways to reduce product cost, and if necessary give back enough of their compensation package to make that work. There a business people lacking high school diplomas who could tell you this; MBAs could tell you this; but technology guys often are mesmerized by their toys and forget they are in business.

SGI has some pretty nice products. Supercomputers. Those should be priced to generate a profit. Graphics software. Why should that not make a profit? Servers - lots of companies make profits selling servers. Professional services - how can you not make profits on professional services? (Technology geniuses have found a way to do that: pay your staff more than you charge the clients! Fail to take overhead into account!)

Rackable's last acquisition did not go well. Management said that RapidScale storage technology would be the Next Big Thing (back in 2006) but it was a failure that apparently could not be sold once Rackable no longer wanted it.

On the bright side, Rackable does have some pretty amazing talent, so maybe management can turn the company around and make hay with the SGI assets too. Rackable got big quickly because it took the need to lower power consumption at data centers seriously. It is still very, very good at that. It has clients like Amazon, Microsoft, Facebook, and Yahoo, and it has been making advances internationally and in new specialties like Defense and Oil exploration.

SGI looked like it should be valuable in the 1990's because it put technology first, before profits. Some times that works, but usually it does not. The best business model wins when the technology is good enough. The best business technology solves enterprise problems at a price that beats the alternatives. Profits have to be part of the model, or else you can't keep up research and development efforts. The leading edge wanders off to ATI or NVIDIA.

Because SGI is selling itself off through a bankruptcy, rather that directly with a merger, it is possible that debts will disappear. This, in itself, might help it move towards profitability, but even if it is the case, can't change the need for better management.

I'm not selling my Rackable stock, but I'm not buying more right now either. I'd like to see what Rackable does with the SGI assets over the next six months or so. If Rackable runs SGI like SGI ran SGI, this merger will sink it.

See also my Rackable Systems page, with links to Rackable analyst conference summaries.