If you check the list of best performing YTD stocks on the Nasdaq, you will come across a name that is familiar to many long-time tech investors: Silicon Graphics, or SGI. The company both is and isn't the Silicon Graphics you're thinking of: SGI was formerly known as Rackable Systems, and it adopted the Silicon Graphics name after buying the old Silicon Graphics out of bankruptcy in May of 2009 for a mere $42M. Although Cramer has suggested that the company made a mistake taking on the Silicon Graphics name because investors remember an equity value destroying bankruptcy, I think he is missing the fact that for many investors the association with the flameout of the old Rackable Systems name would be just as bad or worse. The original Rackable Systems business focuses on data center infrastructure such as rack-mounted servers and storage while the old Silicon Graphics business provides high end technical computing products such as supercomputing clusters. The company's stock has been on a tear, up almost 150% YTD and over 35% since its earnings report May 3rd. Given the history of the company and the lack of profitability, I set out to figure out why it has become a momentum investors' favorite. It's a bit of a head-scratcher, but as best as I can figure whoever is bidding this stock up is not looking very deep under the covers, and doesn't really know what they own.

If you listen to Silicon Graphics conference call you will hear a lot about cloud computing infrastructure, which is not surprising given investor focus on the space. While SGI does provide cloud computing infrastructure, it only does so in the same sense that companies like Dell, HP, IBM and Super Micro Computer do. That is, they provide the commoditized server and storage hardware used in big Internet data centers, a low-margin and competitive business. The difficult nature of this market is why the old Rackable never worked out like investors hoped, and the reason that the company reported scant 22% gross margins on product sales in its most recent quarter. The company does have some storage business but this also seems to have little competitive differentiation and includes reselling the commodity hardware line provided by LSI's Engenio (which is being purchased by NetApp). Amazon is a large customer at about 12 % of the last nine months' sales, but for some reason has been reducing purchases with SGI. Sales to Amazon in the first nine months of the fiscal year were down over 21% from the prior fiscal year, and this is while Amazon is growing capital expenditures at over 150%. From what I can tell there is nothing particularly differentiated about SGI's Internet data center server and storage products.

 

The old Silicon Graphics part of the company (roughly half of sales) provides the clustered supercomputers used for high intensity computing tasks such as modeling weather systems or nuclear explosions. While this business is somewhat higher margin due to the recurring service contracts sold with the hardware, it is also low growth and highly dependent on public funding as the primary customers are government, educational or other non-profit laboratories. This is why "public sector" sales were about 48% of revenue in the most recent quarter. This is also not a terribly fertile market for growth and profits as competition against the likes of HP, IBM and CRAY is fierce, orders are lumpy and government budgets everywhere in the world are under pressure. Hence, the old Silicon Graphics' bankruptcy.

 

Given the fact that SGI addresses low margin single digit growth markets with large entrenched competitors, I find it hard to believe that SGI's product lineup or technology is driving the stock. What else could it be?

 

According to financial press releases SGI has raised revenue guidance twice this year, the first time with earnings on February 2nd and the second time March 9th. The first time the company raised fiscal 2011 non-GAAP revenue guidance from $550 to $575M to $570M to $595M and the second time to $600 to $625M. Judging from these press releases, you'd think SGI products were selling like hotcakes and the company was on a rapid revenue ramp.

 

However, there is much less to this revenue growth than meets the eye. It is important to note that looking at SGI's GAAP financial statements gives you a distorted picture, since due to a change in accounting standards it is now recognizing more revenue up front. To their credit the company helps investors normalize for the change in accounting standards by providing a "non-GAAP" revenue reconciliation. Upon closer examination, the company set a very low revenue growth bar befitting its low growth market and merely stepped over it. Original fiscal 2010 guidance called for 7% revenue growth at the mid-point, and the new guidance in early 2011 simply raised the growth estimate from a pedestrian 7% to a mediocre 11%- similar to the revenue growth that old-line tech companies such as Ingram Micro are experiencing. The second revenue guidance raise appears to be solely due to buying out its minority-owned Japanese joint venture (SGI Japan), which will contribute revenue for the last four months of fiscal 2011 and the entire fiscal year 2012. From what I can figure SGI Japan accounted for the entirety of the March 9th revenue guidance raise for both fiscal 2011 and fiscal 2012. Since SGI bought a business doing about $70M-$80M per year in revenue for less than $20M, I find it unlikely SGI Japan is a highly profitable or substantially growing business.

Looking closely at SGI's most recently reported March 2011 quarter provides an even more sobering view of its growth. According to the press release non-GAAP year-over-year revenue growth was a mere 5%. Sequentially, revenue was down a whopping 27% quarter-over-quarter. However, even these poor growth numbers overstate SGI's actual organic revenue growth. The company's acquisition of SGI Japan closed on March 9th and contributed $5.4M of revenue in the March quarter according to the 10-Q, so organic revenue growth was only 1% year-over-year, and negative 30% quarter-over-quarter. Clearly, organic revenue growth slowed down from a strong December quarter, and is indeed barely positive year-over-year. This is not surprising due to the slow growth nature of SGI's markets and the fact that almost half of its revenue comes from the budget-pressured public sector. Not exactly the usual growth profile of a momentum stock like RVBD or APKT. If momentum investors are bidding up SGI on its revenue growth they are either (1) not even doing a cursory job of looking at the reasons for the revenue guidance raises and underlying organic growth; or (2) algorithmic computers who are only looking at GAAP numbers and not normalizing for the accounting change.

 

Surely the market could not be missing something so obvious, so it must be SGI's fabulous earnings, right?

 

If you look at SGI's earnings history on Yahoo finance, the stock price action starts to make a little more sense. After all, analysts only forecast $0.01 of earnings in the December 2010 quarter while the company put up a whopping $0.44 in non-GAAP EPS, while in March analysts expected the company to lose $0.12 only to be surprised by a $0.07 profit. Clearly the magnitude of these EPS "beats" have a lot to do with the excitement around the stock. However, just like with the revenue guidance raises, the EPS upsides come from stepping over an extremely low bar and there is much less to them than meets the eye.

 

First, you'll notice that original analyst expectations were for the company to lose money. Usually companies in slow growth mature markets do not lose money, but since SGI is a combination of two troubled companies (one recently in bankruptcy) this is somewhat understandable. So the conversion to "pro forma" profits represents a real improvement in the business, to management's credit. However, the improvement has only been from making losses to roughly breaking even, and largely driven by cost cuts, so it is not clear that management will be able to take SGI to a substantially profitable high margin position. SGI releases a very tortured set of "non-GAAP" numbers, where they have considerable discretion on what they count as "one time" expenses or benefits. Theoretically, "non-GAAP" numbers should move the earnings close to the actual cash flow per share enjoyed by the shareholders, while in SGI's case they do not. SGI claims to have earned $0.45 in total "non-GAAP" earnings per share in the first three quarters of fiscal 2011, but on a GAAP basis they lost $0.30, and had $15M in negative free cash flow.

 

Once again, digging into the March 2011 numbers shows that there are much less to SGI's apparent earnings beats than meets the eye. If you look at the "Net Income/(Loss) From Continuing Operations" line, you will observe that on a non-GAAP basis SGI went from losing $10.7M in the March 2010 quarter to making $2.3M in March of 2011. There was an actual improvement in the business to be sure, but at every juncture the improvement is less than a casual reading of the press release would reveal. For example, gross margin supposedly improved from 27.5% to 29.9%, a substantial 240bps increase year-on-year. However, a close reading of the 10-Q reveals that SGI had a $1.3M obsolete inventory charge in the March 2010 quarter which was not adjusted for in non-GAAP results, exaggerating the apparent margin improvement by over 100 bps. Similarly, in the March 2011 quarter, SGI's operating expenses were benefited by a one-time collection of a previously written-off bad debt expense of $1.1M, which was also not backed out in "non-GAAP" results. Finally, the biggest swing in the profitability came from having positive $2.5M in "Other Income/(Expense)" in 2011 as opposed to negative $2.5M in the prior year, which basically consists of foreign exchange gains and losses over which management has virtually no control. If you add it all up, of the $13.1M improvement in net income over half of it came from the aforementioned one time items ($1.3M obsolete inventory charge + $1.1M bad debt recovery + $5 swing in foreign exchange).

 

Needless to say, none of this is the basis for sustainable earnings growth and adjusting for the one time debt collection benefit the company lost money on an operating basis. SGI management seems to believe that "non-GAAP" earnings means including stuff that makes them look good but ignoring stuff that makes them look bad. Once again, this company is not GAAP profitable or cash flow positive. Especially odd on the most recent conference call was the CEO's comment that "We grew our cash 20% to $133.8 million and generated $13.7 million of cash from operations. I note this is a 10% free cash flow yield over revenue." Back on planet Earth, the only reason that SGI had positive cash flow during its fiscal third quarter was accounts receivable runoff as a result of its large contraction in revenue, while for the first 9 months of the year (and likely for the full fiscal year) the company is substantially free cash flow negative. SGI's portrayal of its financial results seems to be a consistent pattern of disingenuous positive spin.

 

After digging into the reasons for SGI's massive stock price appreciation this year, I've come to the conclusion that SGI is not an emerging technology story, a growth story or even a value stock with substantial earnings. Rather, it seems like fast money investors who look only at a company's press releases and "pro forma" earnings against analyst estimates have bid it up to unsustainable highs because they don't understand what this company is or what it does. While SGI's management has done an admirable job of taking two troubled companies and getting them close to breakeven, they are going to have to offer investors something much more to maintain this stock price, much less continued appreciation. SGI competes in the widely disdained, old-line IT hardware market. Once the momentum wears off, from what I can tell SGI offers neither the exciting growth story to attract momentum investors nor a low multiple of cash flow to attract value investors.
SGI's forward non-GAAP earnings multiple is at a nosebleed 44x level more befitting Salesforce.com or its customer Amazon. The stocks of its actual competitors such as DELL, HPQ and IBM trade at forward EPS multiples of 7 to 12x. Unlike SGI, the aforementioned companies actually have the cash flow to back up their interpretation of "non-GAAP" earnings. Other small public companies that sell old-line IT hardware at sub 35% gross margins such as SMCI, CRAY or XRTX trade at low multiples of earnings, book value and/or EBITDA. It is an amazing feat indeed that SGI has managed to capture the fancy of speculative investors given its low growth business profile. At 10x forward non-GAAP earnings (where its comps tend to trade) of $0.59 SGI would trade at $5.90 per share, so it is going to have to "beat the number" by a pretty substantial amount going forward to maintain this lofty stock price. Given the difficulties both the old Rackable and old Silicon Graphics had in the past and the company's reliance on the budget-pressured public sector, I think SGI is a compelling short candidate and its price could crash to below $10.00 on a light revenue quarter or even investors simply coming to their senses.

 

Disclosure: Short SGI

Revenue/earnings miss due to issues in Japan or public sector funding.

 

Once momentum turns around, there is not much to offer either growth or value investors.