Can SGI Improve Supercomputer Margins?

Silicon Graphics International (SGI) makes supercomputers. Its clients include government agencies, financial firms, the pharmaceutical industry, Web farms including cloud services, and an increasing number of business verticals. Just before I last wrote about SGI in February it had run into profitability issues [See SGI Challenged by Analysts]. SGI's stock price was at had declined rapidly from near $15.00 per share to under $10.00. Today SGI closed at $9.37. The 52 week low hit $5.02 on May 21st this year. Is SGI ramping its sales and margins, or is the price increase just a dead cat bounce?

SGI is on a fiscal year, but I'll refer to calendar quarters. For the December 2011 quarter revenues were $195.2 million, non-GAAP net income was just $1.3 million.

The March quarter, which is typically sequentially and seasonally lower in the supercomputer business, saw revenues of $199.4 million, with non-GAAP net income of $3.7 million. June quarter revenue was a dismal $179.5 million, with non-GAAP net income of negative $3 million.

So anyone long in the stock has a case to prove.

Market capitalization ended today at $303 million. Even at $180 million per quarter revenue is running $720 million per year. While higher revenues might be good, the real problem this last year has been margins.

Management has noted that their invoicing process for some computer systems had been SNAFU. As a result, they lost money on those individual systems. Sales people were making deals with insufficient supervision, resulting in great steals for the customers, but no profits for the firm. That practice should have ended by now, but because of the size of SGI deals and the lengthy delivery and revenue recognition processes for supercomputers, there will still be some of that low-margin revenue in the September quarter and maybe as late as the March 2013 quarter. In the June quarter SGI rejected several deals in that quarter because of insufficient margins.

Of course rejecting low-margin deals leads to lower revenues, unless the customers are not as price-sensitive as they pretend to be.

Generally, the business background situation is a cause for optimism. The increases in spending on cloud computing, scientific computing, security computing, and high-performance computing (HPC) have been steady and are likely to accelerate in the past few years.

The new SGI UV 2 supercomputer has received good reviews. SGI does not usually break out revenue by type of computer sold, though that would be helpful to the investment community.

The last guidance issued for the September quarter was for revenue between $180 and $195 million and non-GAAP net income in the vicinity of negative $6 million. The quarter is not over, and the macroeconomy has not been helpful.

I should point out that the old Rackable Systems, which bought the bankrupt old SGI to become the current SGI, also had the same margin issues. Management changes from time to time and new management always says they understand the issue and will deliver better margins in the future. Buying SGI stock right now is a bet that the current CEO, Jorge L. Titinger, who joined SGI on February 27, 2012, can break what seems to be an entrenched corporate culture of great engineering combined with disdain for a need to consistently generate solid profits.

The problem may, in part, be competition, in addition to the sad corporate culture. SGI competes with Dell, Cray, IBM, HP, Appro, Oracle and Fujitsu, among others.

Disclaimer: I am long SGI. I will not make any changes in my position for the next week.

Keep diversified!