Seeking Alpha
We cover over 5K calls/quarter
Profile| Send Message|
( followers)

Cray Inc. (NASDAQ:CRAY)

Bank of America Merrill Lynch 2013 Smid Cap Conference

May 7, 2013 10:15 AM ET

Executives

Brian Henry - EVP and CFO

Brian Henry

So Cray is known as the supercomputing company. It has focused on helping our clients so we could really in a sense provide tools, so it helps our customers, solve the grand challenges in science and engineering whether it’s alternate energy, out programs, whether it’s weather forecasting, weather change analysis, new metals, new breakthroughs in health sciences, personalized medicine kinds of things, pretty broad set up of things.

It also applies in oil and gas, a very important application to optimize where the oil will be to maximize return and engineering in many applied ways with one of our customers who, doesn’t want to talk exactly what they do but I like to say that they are doing things with our system that they can’t do any other way that they see things and could model things as they can do never before.

So I’ll just go back on that last slide, just wanted to point out that we really served two markets today and we are leveraging kind of our supercomputing expertise in the Big Data because in a sense the supercomputing is modeling things with lots of data fast and doing lots of complex mathematical operations on the data. We are taking that technology now taking bigger such a data and managing that in Big Data. So, we have two things, we had about $8 billion addressable market in supercomputing, spread roughly half between the Storage products and the supercomputing products and then Big Data about $2 billion but we are working on several things leveraging supercomputer and will talk a little bit about YarcData initiative and graph analytics appliance that we have.

We had a number of the world’s largest computers including currently the number one ranked computer in the world on the benchmark but we focus a lot more on throughput and actual application performance than we do on the benchmarks, not with all of our computers that all exceed petaflops, the petaflops is thousand trillion floating point operations a second so very high performance machines.

Than unlike most clusters where they only run at may be 32 processors maximum, many of our applications run on 150,000 processors certainly more than 32. They can run up to a full machine size in certain applications. We had a number of applications that run over petaflops.

Recent highlights are two big machines, we put in at University of Illinois, our National Science Foundation program system that's referred to as Blue Waters at MCA has a $188 million transactions for Cray and that includes services over time, it is a one and half petaflop machines, biggest x86 machine and fastest in the world and has the fastest storage system in the world that we install that can move a terabyte of data a second and so if you ever backed up your PC disc and you recall how long it takes for you to move one terabyte on your machine if you haven’t had that much it takes hours, one second. So they are very fast, also very large, 25 petabytes of storage and that’s leverages of Sonexion storage product.

Down below is the world’s fastest computer; it’s a mixture of x86 processors and NVIDIA GPU processors, runs at 97 petaflops as machine has been upgraded several times. One of its predecessors at least in cabinet form but different pictures on the cabinet was the world’s fastest computers few years ago.

I tell you that we designed the machine to be upgradable, big user benefit for a customer they were able to swap out fair amount of the insides over a period of time, they are cot effective; have a very-very fast machine.

We differentiate on a number of things and I would say we aren't always the best marketers but the customer experience with Cray is probably very unique and our customers very much appreciate the hands-on, helpful approach that we have but in terms of technology, we do a system interconnect, system interconnect really allows the flow of data through the machine, allows it to scale up to a way we can, it's very cost effective the way we do it, very efficient use of space. We have a lot of softwares, so we have, I think now two thirds or three quarter of our engineers are software engineers, I suppose the hardware engineer, and we have a very intergraded software stack optimized for performance and that is unique most clusters are high enough to shelf software are integrated stacks definitely not optimized for high performance computing typically.

Unidentified Analyst

(Inaudible)

Brian Henry

They are today, yes, no, we have propriety interconnect, we refer to as Aries and a string of others that’s 186 that has both the router and then net, most outer approaches have separate, they separated them out, one big cabinet for switching and then a nick on the chip within the system and the router would be the cabinet.

We have all in one chip if you received in (Inaudible) set up and wiring and compared to Cray we have fraction of the numbers of cables that are required to connect and we have huge advantage mangers in terms scalability performance on big machines. In terms of packaging, we built very dense machines. We custom designed. The board is custom designed, the cooling, the infrastructure, and so we have a cabinet that can put 384 processors in it and one cabinet and special liquid cooling machine. We have these turbo fans that used to blow from the bottom now blow from the sideways through cool air and can leave the machine with the same air temperature coming in as it goes out.

There is lot of cleverness in how we do it to hit the hotspot with the right cooling coefficient. Adaptive Supercomputing vision is kind of Cray, we started this vision some years ago it’s really to have machine like the one at (Inaudible) that has a mixture of different processors and this one hasn’t had GPS and X86 but it can have a different processors and solve the problems.

Each software application has bottlenecks that are solved with different attributes of processors and memory and interconnection et cetera, and we have flexible system in order to it be able to solve this problem for them in a small intelligent way less burden on the application developer.

We spent awful lot on R&D and also many years in the 100 million range and really most of our employees are engineers around the company and what we've done is we've taken the focus we had on the engineering around supercomputers and began to spin it off in the other business opportunities that we’ve had, so we just introduced today the XC30-AC which is air cooled machine. The technical enterprise solution down the bottom can grow what cabinet and little smaller and smaller price point because of that they can go into broader market including universities each year et cetera and into commercial years and we announced two commercial companies that have brought this one in electronics and another one in financial service.

In the financial service app, I can’t talk about a lot let’s say that they think there is some attributes of our system including the Interconnect that they can solve some problems that are very important in their modeling and that they can’t otherwise do.

We have the cluster solutions, we bought a company called Appro we’ll talk about that in a moment that gives us kind of bridge down the range if the market that we can address down to the half million or less level that allows us to sell machines that our customers already buy. They buy Cray to do the big jobs and they have cluster to do development some of the smaller jobs.

This company Appro was number three in the numbers of supercomputer on the top 100, so they sell some descent size machines and they sold a lot of to some of the same customers that we do in the United States and we’re reinforcing that and also taking them international which they weren't. Storage and data management, we talked about (Inaudible) product and we’ll have another slide on that in a moment, very fast, it’s an appliance like thing much easier to do and install in typical storage solutions in our business, which tend to be let`s say computer science project.

Customer engineering is an element that we used our system for a special purposes, we’ll put them in a box somewhere around the world for commercial or research reasons so we just put one up and fit one in a box in a special way did few things and that’s just the packaging we can do some other things in both software and hardware have a history doing that including for US government.

YarcData is an appliance to do data analytics in Big Data and in particular it does graph analytics, I will talk about that in a moment. Sooner all things are leveraging in our technology form. This is kind of our computing a line and it has the clusters in that CS300 series that we announced this morning which is (inaudible) architecture of the XC30 and then we have XC30 big machine there on this line.

Picture of the Sonexion machine and this is a storage machine and in this one box we can put a 1.4 petaflops with 3TB disc so we will soon increase that capacity and it’s really fast, just connect the boxes up. Again, if you have same projects with lot of storage, there would be a massive wires and connections and if any one failed, it doesn’t work which always happen to major project to fix this appliance, they work together reasonably easy and comparatively huge that easier and quite scalable.

And this is what, is the core of many of our systems that we got in hassle last year that we saw including particular MCA’s Blue Water machine.

Big Data, just talked about and few guys here; all the passwords and things like that is; there is a crowd and this is a confused market place. With there a little segmentation, there are many ways to do it but we rolled that up in a kind of full quadrants that we think about different areas; we got kind of the data warehouse and the extensions, some of them are analytic data basis not too many. And that's the Oracle to tera data; the IV MPP2 will be the EMC screen problem, then you will have no sequel data basis which are basically unstructured data basis that are been of all those are in a relatively early stage development.

On the left hand corner, you hear a lot of Hadoop, which was kind of a foundation but that's what Google and actually Yahoo! developed it. And it has been moved on from there by Google but a common way to break up call and to do search, over the Cloud or internally it’s Hadoop.

We think we can bring solutions to the Hadoop map to reduce leveraging our supercomputing expertise that can decrease the footprint required, improve the performance in the cost effectiveness of those solutions today. They are usually just sprawling lower cheap servers that are connected which is fine for some application but for others and particularly if you want it put one in-house in your company we think we have a better solution for that.

And then graph analytics is where YarcData there are some others out there. Many problems when you are trying to solve or actually best to look at in the graph form and some way how they connect, think of the social networks. And for example, we have a healthcare company who was doing research and they can see connections with different kind of DNA and genome data and certain kind of cancers and things like that and they can hypothesize and see links that they can never see before and you really see that, by looking at lots of data than kind of having it sort through in a way that’s different and graph is a little more; I'll talk about in a minute; looking for discovery.

YarcData; well entity we can grade, built on a machine by (inaudible) software stack that we use but built on the machine we look for the intelligence community in the United States which I don’t know for sure, it's classified what exactly they do, but let's just say, it's helpful in following the dots and I think they use it for anti-terrorism and help to save many lives.

So, taking that machine architecture which is very unique that allows a lot of things to go in the memory, unprecedented level that go in to main memory and allows you to interrogate that main memory faster than anything out there by depending on the benchmark 30 to 200 times faster. And so, you can solve problems as you couldn't before we have a database on top of it and no sequel database on top of it and so it’s user-friendly and we have been again working and selling this over this last year almost effectively, they set up a division called YarcData to do it. Unprecedented for our side of entity; Gartner is really excited about what Cray is doing and really excited about graph analytics and they have carved out a space in their role master graph analytics just ahead of the fellow of our Board of Directors and he talked about where he thinks things are going and he definitely thinks graph analytics are the way they're going. Garth also came out with cool product of the year so, and YarcData was named on one of those cool products as well and it's called the Urika System.

So what are we doing that's different. Much to do but now they're saying you're looking for a needle in the haystack, and when you're doing that the needle looks a lot different than the rest of the things, so you can kind of break up the problem into little pieces, use lots of machines and ultimately find that needle or at least be able to search to find that needle.

What we do is more about discovery things not things that you know for sure like you're looking for but things that you think might be true or you're not even sure and you get surprised, so it's kind of like finding a needle in the needle haystack and then there might be different kinds of needles in there as well, and you're looking for patterns and you can't really break up the problem. There are a lot of problems like this. So what we're doing is allowing people to see patterns and then hypothesize so we can take a big load of data in and then start doing queries on things that they never thought they could do and the ability to get results from their hypothesis back which is all part of science right, to hypothesize, test and find that where they can do that over a hugely fast time to see patterns…

So you know we found people that have seen things that they never thought existed were able to run through a bunch of hypotheses that they had and find one or two that had promise that they'd never seen before, so that's the cool stuff that we're doing and it's because we can put so much in main memory, we can load things in and out quickly and then interrogate it fast. It's not just the initial data which are big data warehouses. First you have to get the data in there to change something that takes lots of times with lots of expertise, then you run a query and God help you when you want to run the next one or two with something different, it's really hard so our system is designed to do this well, one of the areas is cyber security course we read about that daily now, in this space it's to look for patterns and threats and you can see those little pictures up there like graph output and you can do all kinds of algorithms to kind of model how you want the graphs to do in terms of degree of tightness and relationships etcetera, they can run a lot of data on this. One of the guys and we've hired a lot of talented people some of the best people in the industry in particular verticals we have a cyber-security guy that used to do it for DIA, used to all the government databases and websites and of course they got tacked a lot. He said it took him 24 hours to run one screen every day and it'd take him two weeks or so to do the first iteration when they had all the false positive and often they didn't do a lot of things because they couldn't they just don't have enough horsepower to find these things and do them so you had to shrink down. Here they could be doing lots and lots of iterations very quickly.

This is a big opportunity for us it's one of the many verticals that are big opportunities for us, other verticals, finance services and a whole bunch of things, health and bioscience and then of course in the applications we're doing supercomputing. Lots of data in weather around the world, for example you could collect all the data and look for patterns in that and see trends and predict data weather in ways that you could never do before.

We did a couple of strategic things last year in addition internally we required of Apple computer for about $25 million, $60 million revenue, company data clusters most notably as I indicated earlier they are number three in top 100 supercomputing list, pretty impressive, ahead of HP, ahead of Dell, ahead of Sun's SGI et cetera, (inaudible). And then we in our story tired a well-known in our industry technical team that helped us implement big storage solution.

Turning to last year, last year was an excellent year we grew from 284 million to 421, some of it's the timing of big deals of course, because to know Cray is to know we have large deals, each quarter is different and can be lumpy in years to years, can be (inaudible). gross margins 36% overall and had 32 million in core operating income and net income is 161, of course we sold one of our divisions to Intel, much of our hardware team to Intel for 139 million net of cost and about a 132 million net. So, it’s a pretty efficient deal thing from our point of view. We ended the year with 323 million of cash, some $8, $9 of cash that was everything went right, we had virtually no receivables, we just did some big deals inventory was relatively small et cetera so that’s a high, we won’t be there unless miracles happens at least for a while and all that happen.

But it’s gives you the sense that some years ago when I came to Cray, we had net debt and some risk at not surviving and today we have depending on how you count it, over 250 million in cash and no debt.

Unidentified Analyst

(Question Inaudible)

Brian Henry

So, guidance for this year, 500 million of revenue which to know that we had a 140 million from the NCSA deal last year and one customer, it tells you that we will good about the core business growing this year and over time what would be the biggest order we’ve had since I have been at Cray and probably one of the biggest effort in Cray’s history and then in the industry’s history.

So, we expect to grow over the prior year, margins in the mid-30s for the year. Operating expenses are high and up and what we’re doing here is the core operating expenses are actually flattish, but we invested in two areas, we invested in big data and we have a huge investment in big data because we think we have an opportunity to create a lot value in that area, so we have the York data initiative and when we have another big data initiative that’s more or lot high uniqueness in attacking the market with Dupe appliance.

And then we have a big investment in the sales force to as we in all these initiatives saw the opportunities to grow and penetrate the market better. So, that’s why operating expenses are going up, we think there will be good returns, we’ve been pretty conservative but now we see the opportunity.

Also expenses gross margins et cetera are impacted by the accounting that we have do for the Appro acquisition. So, we went to non-GAAP reporting for the first time this year and there was about 10 million of non-GAAP, I’d say. Stock compensation is about 50% to 60% of that probably, 60% of that and then the other stuff which that the accountings for the acquisitions majority of the rest.

We’re also doing non-GAAP for tax because we have our large net operating as carried forwards. So our real tax rate that we would pay in cash is probably 7% to 11%, but we’re accounting for that 40% under GAAP at current label from time to time under GAAP you get these huge swings in taxes, in fact we have negative taxes in couple of years as the way accounting works. The non-GAAP picks all that out, and so, actually in the period of loss over an adjustment to the loss carry forward asset account that we have that takes the effect of that out. And so it actually hurts to earnings but we think it’s a better way for investors to look at it. We expect to be profitable for the year on both GAAP and non-GAAP basis.

I’ll just kind of summarize, we have leadership in the high-end power supercomputing. We have technology that we’re leveraging in several gross markets. These markets are growing faster than high performance featured in market and the high performance computing market in itself is growing 8% to 10% a year.

We have an opportunity in big data which is very fast growing marketplace. We have a strong balance sheet.

Any questions?

Question-and-Answer Session

Unidentified Analyst

Just about the operating expenses, your run rate is 40 million over the last year. how long is that think you (Inaudible).

Brian Henry

Well, probably assuming that we’re growing, it will continue at that level overtime some of the non-GAAP stuff related to Appro acquisition disappeared that’s in there but.

Unidentified Analyst

So this ramp up is spending is something that we should anticipate, is this the new level on go forward basis?

Brian Henry

No, but these are in people and resources largely. Well, we probably averaged 50% to 60% as high maybe over 60% in one year to the U.S. government and last year I think it was 52% or something like that. So, we’ve sold a lot to government and we’ve sold to different agencies of the government. So, DoD, DoV, SSI, and Department of Science amongst others, so it’s not in the one pocket.

I think the administration is very focused on supercomputing being a key country competitive weapon, and with cyber security and other threats, I think and just antiterrorism stuff that some of the things we do for department of defense and the intelligence community, are high priorities in fact some purchases, we have the sequestrations, you all know. Yes I would say it has had impact, we have seen deals slowdown and some deals be smaller than we would see but the things that we expected for this year are still, look like are on target and the government has prioritized some of the things that they told us they were going to do even though they have had sequestration so they have told us they are moving forward just recently as a week or two ago on several of the fronts so. we feel good we have good visibility on the opportunities for this year, it is going to come down a little bit to timing but other than that we have very good visibility on where we are right now.

Unidentified Analyst

(Question Inaudible)

Brian Henry

Well yes, so we have university small then we have international that makes up the rest so we have international governments and then commercial last year was over 10% that’s first time. Since I have been at Cray from 2005, we have had commercial business of any significance this one being over 10%. And we have the opportunity again to do that again and our goal would be to sustain and grow that overtime.

Unidentified Analyst

(Question Inaudible)

Brian Henry

It would probably be have to be driven a lot by the Big Data and maybe some of the cluster stuff, because most commercial applications today don’t require near the performance that we do and so they can buy coasters from us because that’s what they buy and they run 16 process or 32 processors.

Unidentified Analyst

(Question Inaudible)

Brian Henry

Yes, no what got us over 10% was a couple of customers Brian, one (inaudible) size machine, and to do exploration we will call that energy field.

Unidentified Analyst

(Question Inaudible)

Brian Henry

Yes, mostly research and it’s what they’re using products thanks for and they have world’s best scientist, world’s best computational engineers helping them. So in a research lab they will have the big machine and then a lot of support services for that and they’re looking at new source of energy, you wouldn’t model a fusion reaction without building the vessel and modeling the reaction inside long before you’d ever build, and spend millions of dollars doing anything like that.

Unidentified Analyst

(Question Inaudible)

Brian Henry

Yes, and it’s vary…

Unidentified Analyst

(Question Inaudible)

Brian Henry

Having Europe, I would say for the most part, there has been times where Asia has been over 15% that relates largely to a big way and here and there, and most notably Korea weather systems, we have had two wins over last five years does not happen any big than that. Most of it did in Europe, last year, and really shipping this year was first time we have had significant systems going into Germany. So we have big machines in Germany and the UK in particular and then elsewhere around most of the Europe, not so much in France. We have a competitor tied into or used to be home by the government not anymore Paul so we don’t tend to have strong performance there. But the other countries are also good.

We can’t typically sell without special permission into an Eastern bloc country and in China and things without special thing. We could sell a weather system in China, we did on one, and somehow we lost but now have the decision with us (inaudible). Other than that we are pretty unrestricted.

Unidentified Analyst

(Inaudible)?

Brian Henry

Might be, after IBM no bid and therefore they didn’t have enough bidders and then miraculously they bid and won. we don't know what happened. So, any other questions (inaudible).

Unidentified Analyst

Obviously your business (inaudible) but it seems like the things are moving in the right direction, and improving with technology better, what changed over the past year versus five years ago?

Brian Henry

That’s a good question, so we have a number of things going on. we are on our kind of third generation of kind of new systems under the new management team. this one has world best interconnect that software stack most dense system and it runs Intel now. So if you, in the past AMD had better processors for high performance computing we used those, but we switched or added Intel just in time so now we run on Intel chips as well as others. And that allows part of the market that we could not previously address open up that puts us on the same playing field with others who might have an Intel buyers in the thing we get the support of Intel.

So that’s a big shift in addition to all the technology advantage it manages that we have. We have four or five years ago we were pretty focused on kind of rebuilding the business at the very high end which is defense wall and we were working on new opportunities over the last kind of two or three years we incubated opportunities in storage the Big Data appliance and mid-range systems and all those things had been rolled out. We were a little concerned to see if there would be good product market matches and when they were we doubled down investments.

So we have a lot momentum our competitors are good but IBM is not as focuses as they traditionally have been it is not clear they have follow on product in one of their lines that they do. There is rumors that they are selling their x86 business and we do not know whether HPC elements to that are included but we largely compete with them on an Intel based platform in many cases today so would break our heart if IBM made that decision. Hewlett-Packard plays in the mid-market of HPC but really at the high end SGI we see from time to time at the high end but their systems are x86 InfinaDyne-based largely Huskers so we compete well with them at the high end of the marketplace. So in a competitive comparative stats I do not think Cray has ever been better. And so we are winning not every deal we never will some people have to buy from others or choose to buy from others but we are very competitive in the marketplace.

Unidentified Analyst

(Question Inaudible)

Brian Henry

I mean we evaluate it. I would say because of swings in working capital and lumpiness of the business we would like to have over 100 but that is something if there were opportunities that overweighed that really we would look at that view point. So we don't need all the cash we have to run day to day operations. we see at as an opportunity to look at things that we couldn’t otherwise look at and see if we can find something that could add more value and just paying it back to the shareholders at some point if we can then we will probably change our mind in what we do today.

Unidentified Analyst

(Question Inaudible)

Brian Henry

Maybe overtime and maybe with mixed changes but it is pretty hard and in the end Intel is a large part of anybody’s building materials and cost and it is an important piece of the solution so we can differentiate in and around that and we do as most of our competitors probably here in the low 20s. But if that in some cases by the way but there may be opportunities to do that we will continue to work on that it is, in that sense a competitive market we are selling to very specific customers and they have constraints and they are smart about making sure the deals are competitive.

Unidentified Analyst

(Question Inaudible)

Brian Henry

Yes so storage has, carries our gross margin our data I am not sure every year higher gross margin more than 2x best of business probably. Services carry higher margins. Other questions? Thank you very much.

Copyright policy: All transcripts on this site are the copyright of Seeking Alpha. However, we view them as an important resource for bloggers and journalists, and are excited to contribute to the democratization of financial information on the Internet. (Until now investors have had to pay thousands of dollars in subscription fees for transcripts.) So our reproduction policy is as follows: You may quote up to 400 words of any transcript on the condition that you attribute the transcript to Seeking Alpha and either link to the original transcript or to www.SeekingAlpha.com. All other use is prohibited.

THE INFORMATION CONTAINED HERE IS A TEXTUAL REPRESENTATION OF THE APPLICABLE COMPANY'S CONFERENCE CALL, CONFERENCE PRESENTATION OR OTHER AUDIO PRESENTATION, AND WHILE EFFORTS ARE MADE TO PROVIDE AN ACCURATE TRANSCRIPTION, THERE MAY BE MATERIAL ERRORS, OMISSIONS, OR INACCURACIES IN THE REPORTING OF THE SUBSTANCE OF THE AUDIO PRESENTATIONS. IN NO WAY DOES SEEKING ALPHA ASSUME ANY RESPONSIBILITY FOR ANY INVESTMENT OR OTHER DECISIONS MADE BASED UPON THE INFORMATION PROVIDED ON THIS WEB SITE OR IN ANY TRANSCRIPT. USERS ARE ADVISED TO REVIEW THE APPLICABLE COMPANY'S AUDIO PRESENTATION ITSELF AND THE APPLICABLE COMPANY'S SEC FILINGS BEFORE MAKING ANY INVESTMENT OR OTHER DECISIONS.

If you have any additional questions about our online transcripts, please contact us at: transcripts@seekingalpha.com. Thank you!

Source: Cray's Management Presents at Bank of America Merrill Lynch 2013 Smid Cap Conference (Transcript)
This Transcript
All Transcripts