Cray still is focused on the super-computer market, specifically the top-of-the line units usually costing over $500k. Super-computers are used by governments, academic institutions, and some commercial entities for solving the worlds' most complex problems. The key market segments are in scientific research (National Science Foundation, NASA), national security (Department of Defense, Department of Energy), earth sciences (for weather and climate modeling), and in computer-aided design (especially for aeronautics and automobile design). Super-computers require special design in both hardware and software to ensure parallel processing with a minimum of data bottlenecks in interconnection. The density and compute power of these systems is pretty amazing. Cray's XE6 can support 192 AMD (AMD) 8 or 12-core Opteron processors, with a total computer power of 20 trillion floating point operations per second!
Super-computing is a growing market. IDC estimates the addressable market in 2009 at about $3.4 billion, and sees it growing about 6.5% per year through 2014. Evolving scientific areas, such as bio-engineering, nanotechnology, clean energy technologies, climate modeling (think global warming), and terrorism prevention all depend on super-computing assets. Cray is also pushing its Custom Engineering services, where the firm utilizes its significant know-how to develop specialized high performance platforms for discrete tasks. The company has seen solid revenue growth over the past 5 years (about 11% annualized), and I don't think revenue growth will be a huge problem here.
Another positive is the limited amount of competition. Super-computers are a relatively small niche market, with barriers to entry involving customer relationships and technical know-how. Cray's main competitors are the big defense contractors and IBM (IBM). Cray has by far the most powerful brand, a real competitive advantage.
Finally, Cray is in great financial shape. Against no debt, the company holds $132 million in cash, or about 73% of current market capitalization. This contributes to a phenomenally low valuation against enterprise value. Magic Formula earnings yield is insanely high at 56%!
All this said, I don't see Cray as a particularly attractive Magic Formula stock. Let's take a look at several risks:
1. Cray relies heavily on the federal government. Historically, over 70% of revenue has come from the feds (in 2010, it was a bit lower at 62%). The federal government is easily the largest buyer of supercomputing equipment, and will likely continue to be into the near future. Additionally, Cray relies on R&D milestone payments from DARPA's High Productivity Computing Systems initiative. Without these, R&D spending would be about double what is reported now. The DARPA payments end in 2012, and were cut back by 17% after a milestone was missed a few years ago. With the federal government in serious diet mode, this is a clear risk.
2. Cray is not generally profitable. The firm has recorded just 2 years of net profitability since 2003. Sales depend on just a handful of system sales every quarter, making results lumpy and unpredictable. When you combine historic unprofitability with the DARPA R&D credit situation, predicting future profitability becomes a dicey proposition.
3. High performance systems continue to converge towards "off-the-shelf" server components. As commodity computing parts get ever more powerful, they can take over for operations that once required specialized low-end super-computers. This cuts into the lower end of Cray's addressable market.
I do think that Cray is becoming a more sustainable company. 2011 will mark their second straight year of profitability, and there are good market trends at the high end. However, the reliance on the feds and the firm's historic unprofitability are big risks. My analysis comes up with a price target of about $5.75 for Cray, too close to current levels to recommend a buy.
Steve owns no position in any stocks discussed in this article