SGI, with over $600 million in revenues, is a technical computing firm that delivers high performance computing, or what the layman calls supercomputing. The firm sells server, storage, data centre and cloud computing hardware and services. During a visit to Bangalore, Mark Barrenechea, president and CEO of SGI, takes time off to tell Goutam Das why he thinks the firm could succeed in a market that is considered a tricky terrain by many. Excerpts:
Is technical computing a great business to be in? Many companies say that high performance computing is a nice thing to have in your portfolio, but it may not get you great revenues or profits?
There are couple of big things converging. Memory prices have fallen while core counts are exploding. This is going to create a whole new era where everything should be simulated in a computer. Second, you now have big data requirements. Third, with the explosion of cores and decrease in memory pricing, companies are re-looking at computational analysis. So we believe it is a good time to be in technical computing.
I think if you build a company that is only focused in the academic world, it will be very hard to have earnings power. But if you have a balanced business of academia and research, commercial and the broader public sector, then you can grow and make money. We are an example of that. Half of our business is in the public sector; the other half is in commercial.
We define public sector as defence and intelligence, civilian agencies, national tech centres, research organisations as well as universities. I think the failure of high performance computing companies has been when they focused too heavily on just the academic sector. Second, people make dumb decisions. Sometimes you will chase a deal that has no economic value, is high risk technology and low margin. If one or two of these things fail, you can put your company at risk. That?s not SGI. You can create a world class public company focused on high performance computing.
Two third of your business is hardware. Globally, we have seen most hardware companies tilting towards services because hardware is becoming a low margin game. What is your outlook for hardware spending over the next few years?
You have to be differentiated in hardware. If you sell commodity servers, it is going to be a low margin business. You have to be at one end of the spectrum. Margin equals value. Or you have to have high volume on a commodity space. You can differentiate today and for a long period of time. In terms of the outlook, there are a lot of views. The high performance computing market overall is growing at 8% per year right now. All servers combined, is growing 3-4%. We are looking to grow up to 24%? three times the market.
Do you see the cloud as a big demand driver for your hardware or is it too much hype at the moment?
I would say roughly, a third of our revenues come from high performance computing hardware; third comes from cloud hardware and another third from services. Most of the services are driven by high performance computing services. So that makes it two-thirds from high performance computing. Look at the basic premise of a cloud. The purpose of a cloud is to drive high standardisation and utilisition. I have multiple resources sharing the environment; may be, I virtualise the environment. These are good fundamental IT principles. We have a nice name for it today called cloud computing. Earlier, it was called utility computing; before that it was called on-demand. The concepts have existed but to really scale it and to institutionalise it is certainly the latest wave. What?s driving our growth is a variety of things. Certainly, the cloud. We power Amazon and as Amazon grows, we grow. We have been deploying a lot of private cloud as well. So cloud is real, it is not hype.
In the high performance computing market, there must be strong competition from IBM, HP and Dell. How do you access the landscape?
The competition is faltering. Oracle has exited the market. HP is formidable but what I am hearing from HP customers is that they are a bit confused right now. Will HP be in the high-end server business? I think we have an opportunity to offer some trust to customers which HP cannot. IBM is a competitor but they are at the very high end of the market and out next generation product are going to challenge IBM in that sector. And then you have Dell. It is not an innovator; they are a commodity volume provider. It is going to get tougher for Dell as the technology gets tougher. The competitive environment today is favouring SGI.
How do you see the business opportunity in India? Will it be a big market for technical computing?
The Asia Pacific as a region is about $200 million for us. Japan is the largest contributor. The second is Australia. China and India are tied in our business. The Asia Pacific business is bigger than Europe for us. So the region is really important. As we are trying to hit our first billion in revenues, we are trying multiple paths to grow?new product introductions, multiple industries, multiple geographies as well as continuing to acquire.
In relation to India, we have high aspirations. We have a relatively small team here today, roughly 20 people. We think we can grow three times in the coming years, both in the traditional government and research markets as well as in the commercial sector. There is growing awareness in the commercial sector that you can benefit from high performance computing. There is also a growing debate around large scale government funding to step up India into the range of petaflop computing. Will that be a $100 million investment by the government? Or $200 million in the coming years? It is the right time for us to focus more on India based on growing investment.