InformationWeek: Moving from Alpha and Mips chips to Intel Itanium chips for all your high-end servers is a bold move that will have a significant impact on your customers' technology road maps. Is this being driven primarily by financial or technology considerations?
It's a technology decision. It allows us to have a single architecture all the way through the entire product family. When you think about the future of the Web, you know you'll have different platforms. You'll have very large data stores, up to 100 terabytes; high-performance simulation capabilities; application servers; and edge-of-the-Net servers.
Having a single architecture underneath makes all the sense in the world.
InformationWeek: At one point, Compaq had already decided to port Tru64 Unix onto Intel's 64-bit chips, then you backed off that plan. Now you've come full circle again. When did you make this latest decision to go with Intel?
The move to do this was part of an ongoing dialog around architectures that we've been having for some time. There is no question that the advances that Intel has made have made the decision more compelling. Two years ago if you look at where the Itanium family was vs. today, you have to take into consideration the progress that's been made around Itanium.
InformationWeek: Are you concerned that this decision makes you just another box builder that has to compete solely on price?
The real question is, where does our differentiation lie? As the progress line between our 64-bit technology and Intel's drew closer, it became a logical move to go with Intel.
InformationWeek: What's going to drive buyers to 64-bit systems?
We've been most able to take advantage of 64-bit technology in applications that are able to use it--and that's in very large cache applications. For example, in the supercomputer world where we've had a string of great successes, such as Pittsburgh Supercomputing, Celera Genomics, Japan Atomic Energy Commission, and the Australian Supercomputing Institute. Most applications until now have had less of a need for 64-bit architectures. We've been primarily living in a 32-bit application world. As you get into the next generation of applications, the belief is that we will have much more complex data--streaming audio and streaming video. Ultimately, you need larger and larger cache sizes. Much of the future of the Web will be about distributed content, which will increase the need for 64 bit. This is really a question more about the market demand generated by content delivery.
InformationWeek: If the market for 64-bit is exploding, why cede that business to Intel?
What this really says is that Intel has great capabilities and are a world-class provider of chips. They have a huge research arm, they have huge engineering capabilities. So the question is, what is in the best interest of the company from an engineering and go-to-market point of view? There is also a message from a financial point of view: Does making chips make sense in terms of where we create the most value? Is it what we do best? Is a company going to try to compete with Intel just to try to produce the chip? That's not a great way to differentiate yourself. At the end of the day, nobody is going to be able to do it all--develop world-class systems engineering, develop world-class applications, world-class services, and world-class engineering on chips. That's an illogical premise.
InformationWeek: Except for the decision to outsource chips, your strategy sounds similar to what Lou Gerstner has done to restore profitability to IBM. Were you inspired by that at all?
Definitely yes. As we look at the IBM model, and as we look at what customers are looking for--the platforms themselves, the standard building blocks--these are the things that you will be able to get from partners. On top of that, you differentiate with system engineering, memory management, caching, and all the technical services around it, as well as the ability to do interoperability and to offer a solution and services arm to deliver it. But this is not just about becoming a services company; that is not at all our intention. Our combination is to blend superior systems engineering and services, so it results in the kinds of platforms that we think are going to be the future of the Internet.
InformationWeek: Of course, Compaq's biggest problem is that you're losing money--to the tune of $82 million in the first quarter--on your PC business. Compaq still derives most of its revenue from PCs, so how are you going to fix this?
I believe it's still a good business to be in.
You're going to see the drive to lighter, simpler devices. People are going to want to buy access solutions. We sell the iPaq Pocket PC, together with software that allows doctors to manage health care right on the spot. That's attached to a ProLiant server that's attached to the patient care system. That's an example of an overall solution that was actually led by the handheld.
We continue to work on video streaming. We have video-streaming servers which are able to do that, and the next generation is the personal computer with very strong capabilities around stereo speakers and high-resolution graphics. There is going to be a great market for tying the access and the infrastructure together.
Secondly, people will continue to invest in different kinds of innovative access devices. We know that wireless mobility is going to be a whole new opportunity. The explosive growth of the Pocket PC has told us that the wireless world is real. Handhelds are a new category. So we as an industry have to do more creative things to create these solutions where people have reasons to buy a PC other than it just has a faster processor. I'm going to buy a new device because it gives me new kinds of functionality.
But, to answer the fundamental question, we absolutely have to improve the efficiency and the profitability of the PC business. It's still a $20 billion business for us.
InformationWeek: What opportunities do you see in the commercial desktop business?
There is a huge market for buying PCs as a utility. The utility concept is similar to storage on demand and servers on demand. The last thing users want to do is go through this crazy rebid cycle. They want to buy a PC as a utility, pay a monthly user fee. They want us to design the architecture and do the installation, the refresh, and the maintenance, and this allows internal IT shops to go on and do other things.
InformationWeek: Is it possible that Compaq might exit some segments of the PC business?
We are going to be in the access business for sure. It's our heritage, it's fundamental, and I think the integration of the access business is extremely important to us. But I think we have to step up and lead it in terms of developing alternate form factors, in terms of ease of use, and from the standpoint of integration and creative marketing around what the customer wants. The customer is tired of all these PC battles. They want simplicity, and we can help define that. We actually can be one of the ones that define the next generation.
I think the access business is going to continue to be a very strong business. People will continue to buy in new and creative ways, and when we start streaming other forms of content. When we really do streaming audio and video, which is going to come a lot faster than people think, the integration between the infrastructure and the access device is going to become increasingly important.
InformationWeek: People may be tired of the PC battles, but they're probably not tired of the bargains they're getting as PC makers try to slash each other's throats by cutting prices. How can you restore profits in such an environment?
I think there is every reason to believe that prices are going to continue to drop. That is the nature of this whole business. The next battle on price/performance is going to be done through engineering: simpler devices and fewer components on devices that are legacy-free. The trick is to engineer different kinds of solutions.
Take, for example, the iPaq Pocket PC. There is a real lesson not just in the device, which was the next generation in the convergence of the home and business. You can do your E-mail on it while listening to music. There are signs that people will pay for improved functionality or to be able to do something new. They're not going to pay simply to do more of what they are doing today.
InformationWeek: In addition to exiting the chip business, you're launching a services push. But the Digital Equipment Corp. acquisition failed to make you a services leader and things didn't go smoothly. What's going to be different this time as you buy up services companies?
Within the company we don't even talk about Digital integration anymore. That is long gone. And I'm not going to comment about my predecessors. I have a very simple rule: You work with the culture, not against it. We've managed to very quietly integrate some other companies, and it's gone just fine. We've come a long way. Digital was huge. That was a foundation-changing, wrack-the-rafters, entirely different business model. What we're looking for now is things that naturally fit or extend businesses we're in. We look for service companies that have vertical capabilities that will be natural extensions of what we do.
InformationWeek: How do you go about incorporating the culture of an acquired company into your own, especially when it may be coming from a very different business model?
You have to make very clear to everybody how an acquisition fits into the strategic plan. You have to take some time to understand the company's management team. This is not rocket science; you have to make sure the cultural fit is there. And you have to be able to understand what cultural things need to stay in each of the companies. The culture in our consumer group is not the same as the culture in our Himalayas group, and that's OK. The guy who's working on the next generation of fault tolerance isn't particularly interested in the color of the panel on an MP3 player. We've done a good job of having an overlying set of values and having people work within that.
InformationWeek: I'd like to close by going back to the Intel decision. Basically, you're betting your high-end server business on technology that has virtually no track record. Are you comfortable with that?
When you make partnerships, just like when you make acquisitions, you do them with world-class people. There is so much concentration in our business on this week, this month, this quarter. The companies that will ultimately be successful are the ones that are capable of looking beyond that--though I have to admit that I spend a lot of time these days looking at this week, this month, this quarter.
There is no doubt in my mind that Intel has got great engineering, they've got great management, and they will be successful. Will there be some potential hiccups along the way? Sure there will. One of the things we're very good about when it comes to partnerships is that we understand those ups and downs. There might be a problem on any given day, but our relationship with Intel is absolutely rock solid. They'll help us and we'll help them, and over a longer horizon there is no question in my mind that this is the winning architecture. I'm very comfortable with this.