| 2007 | 2008 | ||||||
| Price: | 27.52 | EPS | |||||
| Shares Out. (in M): | 0 | P/E | |||||
| Market Cap (in $M): | 323 | P/FCF | |||||
| Net Debt (in $M): | 0 | EBIT | 0 | 0 | |||
| TEV (in $M): | 0 | TEV/EBIT | |||||
Sign up for free guest access to view investment idea with a 45 days delay.
SGI represents the opportunity to purchase a technology leader in the high-performance computing (HPC) space at a significantly cheap valuation (3.8x CY07 EBITDA for its operating business). Downside protection comes from its cash cow service business, monetization potential of its large non-core intellectual property portfolio, excess working capital and minority stakes.
SGI recently emerged from a Ch11 pre-pack led by Quadrangle, Watershed and Symphony. The three investors currently own 51% of the equity. SGI was thrown into Ch11 as revenue slid from $3.7bn in 1997 to $520mm in 2006 and fixed costs where not cut rapidly enough. The Ch11 process has allowed SGI to restructure these fixed costs (including significant real estate liabilities) and to refocus the business. The company has emerged as a FCF+, EBITDA+ and growing revenue company. A lack of investor communication and obfuscated financial statements resulting from fresh-start accounting has hidden the company’s true profitability. Additionally, monetization of SGI’s hidden non-core intellectual property will unlock additional cash over the next year that will also help clarify the equity story.
Overview
SGI is a designer and manufacturer of the highest of high-performance computing (HPC) systems. These systems are primarily sold to governments and those affiliated with government projects (e.g. NASA and universities). SGI uses standard components (e.g. processors from Intel) but integrates them using their IP-protected architectures and software that allow their machines to significantly outperform competitors. SGI competes against IBM, Cray, Bull, Sun and HP. SGI’s “signature” system is NASA Ames’s
Financials
SGI is currently trading at 8.8x CY07 EBITDA. However, when I add back a conservative estimate of potential proceeds from the cash monetization of non-core IP (trial date 11/13/07), cash inflows as working capital normalizes and the value of minority shareholdings (SGI Japan), SGI is trading at 3.8x CY07 EBITDA.
Here is my math:
SGI’s Q2’07 (12/29/06) balance sheet has net cash of $13mm: $109mm of cash (including equivalents and restricted), $86mm of debt and $10mm of cash liability related to bankruptcy. The company has 11.25mm outstanding shares. The first batch of management options and restricted stock was issued on 12/1/06 (options priced @ $18.03). This issuance added the equivalent of .47mm shares (using treasury method) at today’s stock price. At $27.52/share, the fully diluted market cap is $323mm and the enterprise value is $309mm.
SGI’s debt facility was negotiated by management and equity investors together when in Chapter 11. The EBITDA coverage covenant was set during this time. I feel comfortable that this covenant is achievable since it was set by investors, including one that also has a private equity practice which has a lot of experience setting covenants (i.e. Quadrangle). The CY07 EBITDA covenant is $31.2mm. Typically these covenants are negotiated with a “cushion” that allows the company error in its business plan. The typical cushion on the last few similar debt facility that I’ve seen is ~15% on the low-end. Therefore, I believe that $31.2mm * 115% = $35.8mm is roughly the budget that the inside investors and management shared with the bankers (which is probably a little below their own budget – you don’t want to miss the bankers’ internal projections either!). EV of $309mm / $35.8mm CY07 EBITDA = 8.8x.
SGI’s working capital is slightly inflated using pre-bankruptcy metrics (average of 2002-2005 metrics, not including the distressed 3 quarters prior to bankruptcy) for balance sheet items. This should free up ~$10mm of cash as the metrics return to their averages. However, over the short term (the next 2 quarters), working capital will actually get worse as the company prepares the delivery of supercomputers to a significant client in Germany and builds inventory ahead of delivery.
SGI holds 10% in SGI Japan, a computer wholesaler in
SGI has over 700 patents. The majority of these patents are not related to SGI’s current core business. They are generally related to SGI’s old legacy business and were assembled back when SGI had a huge R&D budget. This portfolio was valued at ~$40mm at the midpoint of a range by CRA during the bankruptcy as part of a liquidation analysis. These exercises are routinely low-balled in a pre-pack as the guiding party leading the pre-pack wants to minimize enterprise value in order to encourage increased ownership for their cash infusion. One potentially valuable patent from SGI is Patent No 6,650,327 for which SGI filed a patent infringement lawsuit against ATI (acquired by AMD) on 10/23/06 (1 week after SGI emerged from bankruptcy). For those who want to understand the tech speak behind the patent, the comments on the techreport.com blog are very useful. It is also notable that ATI’s competitors (i.e. Nvidia and Microsoft) currently do license this patent from SGI. Microsoft got the license in a large cross-licensing deal. Nvidia got the license as part of a settlement to an SGI filed suit in 1999. While the suit’s settlement amount has never been publicly announced, industry estimates seem to be in the range of $50mm+ which was based, at the time, on expected revenue from graphics cards incorporating the patent’s technology. Since this time, Nvidia has grown from a $500mm EV company to $12.5bn along with the rest of the graphics industry. ATI was acquired by Intel in July for $5.4bn. This implies that a settlement today would be significantly higher, obscenely higher… The trial is set for 11/13/07. The docket can be accessed at http://pacer.wiwd.uscourts.gov/ using case number 06-C-0611-C.
The other large area of potential litigation is around SGI’s LCD display patents. A search on the
Together these two opportunities can clearly be a very large number for SGI. It’s easy to make the case that proceeds from these patents alone should be significantly above SGI’s current enterprise value. Potential paths to monetize these patents include settlement, trial or acquisition (of either the patent or SGI in its entirety). SGI has taken the first step here by filing suit against ATI.
I feel comfortable that the minimum value received from ATI will most likely be north of the ~$50mm received through the Nvidia settlement in 1999 and, using simple ratios, the settlement can be as high as $500mm+. Additionally, there is potential value from the upcoming LCD display efforts. For purposes here though, let’s use a conservative risk-adjusted minimum value of $150mm of proceeds for the monetization of these non-core patents. Backing this value out lowers enterprise value to $174mm and EV/CY07 EBITDA to 3.8x.
So is SGI’s remaining business worth $135mm? I’m a value guy so unfortunately I don’t have a perspective on SGI’s ability to grow the business and for SGI to ultimately get valued in line with comps (>1x revenue on ~$500mm of revenue) for which the growth guys will try to sell you on. However, I do see value in a couple of their revenue streams. The easiest one is their service revenue stream. Their storage product revenue can potentially also be valuable.
Service revenue using Q207 run-rate is an annualized $233mm and has a 47% gross margin. The company calls this their “cash cow”. The disclosure statement plan assumes 10% churn – an assumption for which management is still “comfortable”. At a 12% discount rate and taxed at 35%, the DCF of the gross profit would be $320mm. While SGI spends this gross profit on S&M and R&D across the rest of its business, an acquirer of this service revenue would not. For an HP or IBM, this service revenue would largely fall to the bottom line. This alone more than covers any downside to the equity and provides a very healthy margin of safety (67%). Also note that I used a 35% tax rate; SGI post-Ch 11 has $8.1mm of NOLs/year through 2020– this is worth an additional NPV of $41mm assuming a 12% discount rate. SGI as noted in SGI’s recent 10Q, they are in discussions with the IRS and Canadian Revenue Authority on additional tax issues that can provide more tax credits.
An additional pocket of value may be SGI’s storage product revenue. Tucked into the product revenue line, these products have been sold more as an add-on to HPC sales rather than standalone. Without any dedicated salespeople, SGI has sold $24mm of storage products over the past 2 quarters ($48mm/year annualized). Management indicates that they are adding S&M resources to finally sell and market these products to customers. My diligence calls have indicated that these products would have substantial potential with committed salespeople or additional resources. This business competes with NTAP (trading at 3x revenue, 20x EBITDA according to Capital IQ) and ISLN (20x sales). Management has indicated that they are redirecting resources to start selling this product. If this division grows, I would not be surprised to see it spun off at a later time.
One additional point on the financials since it will likely come up once you dig in – my interpretation of SGI’s latest 2 quarters. While the company provides EBITDA, they do count some 1-time costs that went away in January 2007 and won’t be in future quarters. For Q1, this included $6.5mm of costs (audit fees + old HQ costs (they moved on 1/1/07)) and for Q2, this included $1.5mm of costs (much lower audit fees + old HQ costs). Using this addback, my EBITDA number for SGI’s FYQ2 is $5.6mm. Generally, I think SGI’s results were positive for these quarters. The company grew revenue and backlog coming out of bankruptcy, which are big question marks for any technology company emerging from bankruptcy. As the company can now begin selling customers again and as costs continue to decrease, EBITDA will increase dramatically by the end of the CY. With Quadrangle and Watershed on the board as well as management’s verbal indication that they are intent on taking more costs out of the business, I think it is safe to assume that there is further upside on the cost cutting potential. The post-bankruptcy revenue pop (in a couple quarters from now as the sales pipeline refills) and cost cuts should easily get us to covenant EBITDA for CY07.
FYI - a useful resource to understand the accounting addbacks (SGI’s accounting has a ton of adjustments due to the bankruptcy process) are “pro-formas” that SGI provides on their website:
http://www.sgi.com/company_info/investors/statements/fy07_proforma.xls
http://www.sgi.com/company_info/investors/statements/fy06_proforma.xls
Comparable valuations:
Comparable businesses include CRAY, HPQ, DELL, SUNW and IBM. Cray is the most comparable in being largely a pure-play HPC company. These companies trade at 0.8-1.7x forward revenue (Cray, the most direct comp, trades at 1.2x). They also trade at 8.5x-13.3x forward EBITDA.
Conclusion:
Price target: $40-$47 based on 8-10x CY’07 EBITDA + $150mm for non-core IP + $10mm of excess W/C + $14mm for SGI Japan. Alternatively for super-bulls, revenue multiples (perhaps which should be applied for a company with depressed margins with a potential M&A event) would imply a higher price target. For example, a revenue multiple range of 0.8-1.7x + $150mm for non-core IP + $10mm of excess W/C + $14mm for SGI Japan would imply a price target range of $53-$94. Using Cray’s revenue multiple would imply a price of $71.
Here’s another way to look at the opportunity which I sometimes find useful: at today’s current price (and backing out the excess W/C and SGI Japan), you are able to buy SGI for 8x CY’07 EBITDA or 0.53x run-rate revenue. This valuation is below comparable valuations so you have a little upside / margin-of-safety here even without improving financials. Moreover, you would be getting a free potentially HUGE option on SGI’s ability to monetize its significant non-core intellectual property as well as an option on a potential hidden-gem storage business.
Positives:
Negative:
Catalysts:
| show sort by |
Are you sure you want to close this position Silicon Graphics?
By closing position, I’m notifying VIC Members that at today’s market price, I no longer am recommending this position.
Are you sure you want to Flag this idea Silicon Graphics for removal?
Flagging an idea indicates that the idea does not meet the standards of the club and you believe it should be removed from the site. Once a threshold has been reached the idea will be removed.
You currently do not have message posting privilages, there are 1 way you can get the privilage.
Apply for or reactivate your full membership
You can apply for full membership by submitting an investment idea of your own. Or if you are in reactivation status, you need to reactivate your full membership.
What is wrong with message, "".