Investing in Supercomputers with Cray (CRAY)
In a recent article, we talked about the difference between supercomputers and mainframes. We were quite surprised to see how pervasive IBM mainframes are in the world’s biggest companies, and even more surprised to see that they offer unmatched performance, like being able to encrypt data 18X faster than x86 platforms at just 5% of the cost. With yesterday’s news of IBM reaching new milestones in quantum computing, you’ll want to read our article on Quantum Computers vs. Supercomputers vs. Mainframes. In that article, we pondered as to how shares of supercomputer maker Cray Inc. (NASDAQ:CRAY) would react to news of quantum supremacy. The answer is a function of how supercomputers like those built by Cray are displaced by quantum computers. We can speculate all day about that, but what we’re interested in learning about is Cray today.
Founded way back in 1972, Seattle based company Cray Inc. is a publicly traded company with a market cap of $746 million and 2016 revenues of $629 million. You could write a pretty engaging novel about the history of this company, but in the absence of such a book you might read this interesting article published by CNN Money back in 2003 which describes some of their volatile history. If you were a shareholder in CRAY since their IPO back in 1995, you’d have lost about -12% of your money compared to a +539% return of the NASDAQ over the same time frame:
As with any tech stock, Cray shares are volatile, but not in a good “NVIDIA volatile” sort of way. Forgetting about past performance, what we’re interested in is what their business is getting up to today. We tried listening to their last earning call, but it was unusually painful. We couldn’t take any more of the “great question Aaron, we’re really excited about… uh.. about this” responses to the questions pitched by analysts during the call so we decided to look at their latest 10-Q filing. The time frame for the following numbers is the first 9 months of 2017.
Revenues are broken down as 52% product and 48% services. While we might be tempted to think that “services” simply involves the maintenance and support of installed supercomputers, it’s actually more complicated than that. Here is a more granular breakdown of where their revenues are coming from:
It makes sense that if you’re selling supercomputers that you’d also try to sell other products and services using those same sales channels. While the increase in +150% in engineering services revenues is promising, the -34% decrease in supercomputing hardware revenues is disappointing. Maybe they’ll end up with some big surprises in Q4-2017 but the earnings call wasn’t overly optimistic.
Just going back to this whole earnings call thing, when are companies going to realize that these earnings calls (much like most meetings in the corporate world) are largely just a waste of everyone’s time and something that could be handled with a few concise emails? Get rid of the “corporate communications department” that probably spends most of their time on Facebook between earnings calls and use those cost savings to improve the bottom line.
While past performance is not indicative of future performance, a lackadaisical share price performance over the past 22 years for a technology company shows that something isn’t right. The current CEO, Peter Ungaro, has been at the helm since 2005. Without trying to oversimplify things here, the way most publicly traded companies operate is that the CEO has to report to a Board of Directors (their boss essentially) and guide the company towards increasing their earnings-per-share over time. During Mr. Ungaro’s 12-year tenure at Cray, the share price has been roughly flat (Dec 31, 2004-today) while the NASDAQ has returned +212% over the same time frame. If Cray can’t achieve results in a bull market, what should we expect when an inevitable recession hits?
We really tried to get more excited about Cray, but kept getting that same sort of feeling you get when reading about Ginni Rometty’s “bold vision” for IBM. A few more red flags popped up in the filings that were worth noting:
- For the first 9 months of 2017, the percentage of total revenues attributed to the U.S. government came in at 64% compared to 47% for all of 2016. Given the volatility of the current U.S. political environment, this seems like a huge risk for the company. Since 27% of their total revenues for 2017 to date were to countries outside the U.S., that implies that just 9% of sales are to U.S. companies, which helps offset the political risk but not by a whole lot.
- In 2015 and 2016, Raytheon brought suits against Cray for a series of patents relating to both hardware and software. While Cray offers up the usual “material losses to these matters may be remote” comment, Raytheon is a $53 billion company and probably has more legal resources than Cray to make things happen. If you want to check the case out in its entirety, here it is.
What’s to like about Cray is that they’re truly a pure-play on the supercomputing theme. If they can find the time to lay out a strategy in the form of an investor presentation, that might be useful for retail investors who don’t want to suffer through painful earnings calls. What’s not to like is their dismal returns and largely sideways trading.
There’s also the promise of AI which has been talked about by Cray a fair amount lately. Just yesterday, they announced that Samsung is picking up a supercomputer for use in research into artificial intelligence including connected cars and autonomous technologies. The day before yesterday, they announced an agreement with Intel to collaborate on a software stack for deep learning on Cray systems. There’s lots happening but this all needs to translate into something meaningful for their bottom line.