OK, 2016 wasn't the best, but look for a buyer? That's Cray

Slowing market sparks ideas... but only in analysts' minds

8 Reg comments Got Tips?

+Comment A poor year for Cray was rescued by its fourth 2017 quarter’s results.

  Q4 2016 Q4 2015 Change Full 2016 Year Full 2015 Year Change
Revenues $346,600,000 $267,500,000 29.6% $629,800,000 $724,700,000 13.1%
  $51,800,000 $20,300,000 155.2% $10,600,000 $27,500,000 61.5%

Full 2016 year revenues went down 13 per cent with net income dropping 61.5 per cent compared to 2015. But the fourth 2016 quarter’s results were good, with revenues rising 30 per cent annually and net income increasing heroically by 155 per cent. These results exceeded what Wall Street analysts were expecting.

For 2016 overall GAAP gross profit margin was 35 per cent; it was 31 per cent a year ago. OPEX in 2016 was $211.1m which was more than 2015’s $184.7m.

Cray CEO and president Peter Ungaro’s canned quote said: “While 2016 wasn’t nearly as strong as we originally targeted, we finished the year well, with the largest revenue quarter in our history and solid cash balances, as well as delivering profitability for the year.” Also 2016 was Cray’s second highest revenue year in its history; not bad then.

What further light does the earnings call cast on this?

Earnings call

CFO Brian Henry said: “Driven by lower product revenue in 2016, total revenue declined by 13 per cent for the year to $630m. Product revenue was $499m compared to $601m in 2015. Service revenue increased by 6 per cent to $130m for 2016. Storage revenue was down by about 20 per cent year-over-year, driven primarily by slower market conditions from our high-end supercomputing solutions, which tend to be a strong driver for storage sales.”

Ungaro’s prepared remarks pointed out: “While it was disappointing to see our revenue growth take a step back in 2016, when you take a longer view of our results, such as over the past few years, you will clearly see that we are delivering strong growth and winning in the market.”

Cray, Ungaro said, is seeing “very slow conditions in supercomputing … with fewer opportunities, both in total numbers and dollars. In fact we believe the high-end of the market which we service … was down by more than 25 per cent on a revenue basis in 2016 and down even more on a bookings basis.” It can not see a rebound coming in this market and so can’t “provide a reasonable range of revenue expectations for the full year of 2017.”

Also “Our cluster business was not strong as we had targeted for 2016,” but “We had another strong quarter in data analytics led by momentum on our new Urika-GX platform.”

Ungaro said the company had a first focus on achieving its outlook for 2016, which it did, and a second one on winning new business, which it did not. The CEO confessed: “While we’ve begun to see some positive signs in customer activity, it has continued to lag our expectations for the past few quarters, primarily driven by reduced bid activity across the board. This has clearly had an impact on our visibility as we turn to 2017.”

Consequently it doesn’t think it will grow revenue in the first quarter, and is suggesting around $55m; that’s a heck of a seasonal drop.

Cray in 2017

For 2017, “we have two main goals … The first is to win new business for this year and beyond, and the second is continued to expand into the commercial and big data [analytics] markets.“ He thinks the high-end, supercomputing market will rebound at some stage, perhaps in 2017. With multi-quarter sales cycles Cray wouldn’t see the results until 2018.

It’s target markets for 2017 are:

  • Big Data
    • Data analytics (with Urika-GX)
    • High-performance storage (Sonnexion arrays and DataWarp IO accelerator)
    • AI market subsets deep and machine learning (XC50 and CS-Storm supercomputers)
  • Commercial - vertical markets
    • Energy
    • Manufacturing
    • Financial services
    • Life sciences

Ungaro says Cray is broadening its product set to be more attractive to these markets.

Cray is finding it hard to predict how it will do in 2017, “due to current market conditions,” as it said in the earnings call. In answer to a question Ungaro said: “In order for us to grow, in the first half of the year, the orders would have to really start to pick up. And so, we haven’t seen that, as I mentioned; we haven’t seen those opportunities that are hitting our pipeline, which our pipeline is remaining pretty strong, but we are just not seeing a progress fast enough through the cycle yet and moving to where our customers are placing orders and we are going to install and get revenue.”

Also, and perhaps tellingly; “opportunities moving through the pipeline are moving slower,” indicating sales cycles are lengthening. Although Cray’s win rates are remaining strong it is having to work harder and longer to win business.


Cray is a big iron company and big data analytics typically uses masses of commodity servers - ie, not big iron. Unless Cray can somehow offer better performance than existing analytics hardware, or somehow get hold of better software, it’s hard to see how it could make much progress. Its Urika-GX product is a high-end multi-processor, 16, 32 or 48 node Xeon system, and competes with systems from HPE (+ SGI), Dell, DDN and others.

Even commodity server vendors are getting in on the act, witness Inspur with Diablo Technologies’ help yesterday.

Enterprise IT suppliers are moving up into commercial/enterprise HPC as pure supercomputing/HPC suppliers move down from their high-end government/academic market into it as well.

Back in the supercomputer area, Cray is hopeful it can win business from customers with old IBM Blue Gene systems upgrading to newer kit using Intel’s Knights Landing or coming Skylake CPUs. However Ungaro doesn’t know when the Blue Geners will jump; “The bigger question … that we just can’t answer for you right now is when are these guys going to make decision? So, they all know that Blue Gene machines are getting pretty long in the tooth right now. But it’s really [about when they make] those decisions and moving forward.”

Supercomputers cost a lot, a real lot, of money. If souped up commodity servers from Inspur, Supermicro and others can do the same or a better job more cost-effectively then Cray has a fight on its hands. Also, in the storage area, DDN is a real feisty competitor, developing its products fast, and Cray is in danger of being outclassed here.

Could Cray be sold?

A hard rain is falling on Cray. Is it in real trouble? What about Cray selling itself to get out of this hard rain area? Here’s Ungaro answering a question about that:

Well, I think from a shareholder perspective, we always have to say -- we always consider alternatives for the business, whether -- regardless of where our stock is trading, high or low. Right now, I think both the management team as well as the Board are very aligned, and that we feel that the long-term prospects of the market look good for the Company.

And our competitive position in the market is solid. I mentioned that the market declined a lot more than our revenue declined from a revenue basis last year. So, we’re doing very good in the market and our win rates are holding up really nicely in the market.

Is Cray hoping for an HPE-SGI takeout for itself? “I don’t know how to even answer that question, I guess that’s always a possibility …. It’s definitely not in our -- it’s not a card that we’re playing or anything.”

So Cray is not looking to be bought but analysts are starting to think an acquisition might be on the cards, possibly looking at Dell as the white knight, maybe even, wild card here, Cisco. Let's throw HPE in the pot as well for three of a kind. ®


Biting the hand that feeds IT © 1998–2020