CRN Interview: Paul Otellini, Intel

Intel President and COO Paul Otellini recently met with Industry Editor Craig Zarley and Senior Editor Edward F. Moltzen and discussed a series of issues related to his company, the channel and the overall industry. Here are portions of that interview:

CRN: You've targeted two areas where you're going to be spending more money this year: R&D and the other was the channel. What's in your thinking? Why does the channel rate as a spending priority?

Otellini: The channel has always been a spending priority. ... For 35 years now, about a third of our business has gone through distribution at one time or another. And that really hasn't changed. And it's a different type of product, different points in time, different countries around the world. But the channel has been from Day Zero an integral part of our business. And our investment in the channel this year really takes on a couple of different aspects. One is, I don't want to use a cliche, but breadth and reach. We are entering into another x-hundred of cities around the world, and going into tier-three, tier-four cities in many countries, and entering for the first time in a lot of countries where we haven't had a presence before with direct, Intel-badged employees, which begins the process of recruiting and getting that whole cycle going, which we've seen work so well and what we'll say are now mature markets. And the other is strengthening up our offerings in the channel to sell a broader offering, a broader representation of Intel's product offering. You saw that initially with desktop products and things like flash memory then you saw the server boards, and now you're seeing it with notebook products. I think that that's just a natural evolution of the fact that not only is Intel moving toward a solution-based kit of products, but our customers and their customers need that for themselves. And I think that's just a natural evolution. And it's important for us that the channel is capable of selling the breadth of Intel's product line. That's the only way they're going to be competitive over time.

CRN: If you're increasing spending this year, where specifically are the additional funds going to go? Is it across the board, or are there targeted areas where you want to spend more money?

Otellini: I'm not going to give more granularity than I gave you. Which is that it's in two areas. It's bringing Intel for the first time into a number of cities, mostly in emerging markets where we haven't been before, and that is a big part of the expansion program. And the other is in products.

CRN: Looking out, realistically, what are the big growth areas that you see for Intel and, by association, for Intel's channel?

Otellini: Our interests are, as we've been describing in the last few years, to leverage this acceleration of convergence that is happening not just to our industry, but to a lot of industries around us. We've identified four focus areas for our product development. One is the digital home, which we've talked about. The other is the digital office, or the enterprise. The third is what we call mobile Internet clients--handsets and handhelds and notebook computers. They're all kind of the same. I spoke at the 3GSM conference last week in Cannes and said there all a phone is is just a small computer. It's not about voice anymore, it's about data and video and other things you do on a computer. There's no reason we shouldn't be involved in those. And then the fourth one is the opportunity in the telecommunications infrastructure, networking infrastructure, which has had a tough few years. But as they come out of that period of overinvestment, I think they are likely to make the next round of investments based upon standards. It's just more cost-effective. And it reflects, I think, a lot of hollowing out of some of those traditional [patterns] that you've seen in the past. I think we've got a good opportunity to sell standard products and they're both Intel Architecture-based products but also Intel network processors inside a blade infrastructure that handle both network applications and packet processing. If you think about all those, is there growth in communications or computing--the answer has to be 'Yes,' because every one of those platforms I described has computing and communications as an integral part of what it is--and we look at these markets along that horizontal slice vs. how many processors will it consume, or how many graphic chips or how many Ethernet chips. We look at it much more as, 'What's the opportunity to sell silicon into all of the platforms?'

CRN: Is there a PC refresh cycle starting?

Otellini: I can't define that any more. We all lived through a wonderful, probably once-in-a-lifetime, few years which was the Internet boom and the euro conversion and the Y2K conversion elsewhere. All three of those drove hardware purchases ahead of their natural curve, in hindsight. You look back at the natural growth curve vs. the accelerated growth curve of the late '90s, and then the corresponding trough of 2000 through 2002, it's very clear there was acceleration. So I think it's healthy now that we're back to normal sort of growth. It's not driven by, from what I can tell, extraneous factors. It's driven by what's important to companies, which is productivity. And what we see is that people didn't stop spending on IT in the downturn. They just shifted their spending. Think about it. They had relatively new equipment from Y2K or the fire sale of the dot-coms. They had a lot of new applications they're bringing in-house. Everybody was moving to SAP and Oracle and PeopleSoft and so forth--CRM systems. And people spent, like us, a couple of years building up systems. Now they're starting to get deployed on them. The best leading indicator for that I could give you on this is we've had six record quarters of server chip sales--consecutive record quarters--but the last year of server chip shipments were very, very good and I think that reflects people deploying these enterprise solutions. Those enterprise solutions will lead to new use models, whether it's mobile computers penetrating the workforce through Wi-Fi and VPN or remote work or outsourcing, that whole trend of, you know, it depends on your politics, the fact that work is happening elsewhere and fiber optics are pulling things together in a collaborative work environment that you need higher degrees of IT capabilities to solve and to deal with. How do you deal with a 24x7 development environment? You have to have better tools or equipment.

CRN: Just looking ahead into your crystal ball, as you're getting ready to assume the role of CEO, what are your top priorities?

Otellini: Whatever [Intel CEO] Craig [Barrett] wants. (Laughter) We're a top-down company. Seriously, I think that we have made an incredible investment the last three years, 2000 to 2003. We invested $16 billion of capital into a newer factory network at a time when people thought initially we were crazy. And now people are saying, 'My God, that was the best bet in the history of computing.' It has given us an unparalleled asset coming out of it. And I think what I would look to do is how to take advantage of that: How do I use that factory network to drive, you know, more product, exciting product, lower costs, better capabilities into the marketplace, and, hopefully, benefit the company.

CRN: One thing you said earlier is the nature of the business going more toward solutions. And that has implications for Intel and for its channel, where it's not the speeds and feeds kind of world and it's a business solutions type of environment. Does that change the nature of your channel, the people you bring into your channel, or how you approach your channel now that it is truly a software-driven, solutions-driven environment, now that you're selling the fastest product?

Otellini: It does, but that change started happening five years ago. We really were looking out at the environment for enterprise computing in the late '90s and this was the era of Sun being "the dot in dot-com." And what Sun had really done exceptionally well was build what we would now call an ecosystem around them, whether it was Java-based, or application based, or collaboration with Oracle, BEA, all that other stuff. They had a very good total solution that was not easily replicated from the PC side of the world. Certainly, there were enterprise computing companies that provided similar types of things, but Sun had a lot of the momentum. So what we did--as we analyzed how they did their solutions and how they built their ecosystems--we started replicating a lot of that. And I think much of what you see of Intel today--in the channel, in our go-to-market capability, in how we sell to enterprises, how we address small business--is the manifestation of that. We've got thousands of software engineers helping enable solutions, optimize solution stacks on Intel. We don't sell application software. But, yet, we have thousands of engineers helping other people tune their software, move it over and get a complete stack. Nowhere is that more true than Itanium. What good is an Itanium server unless you've got a stack that runs the entire SAP or the entire Oracle or the entire PeopleSoft or whatever you have to be running? So I think, if anything, the channel now benefits from a lot of that heavy lifting. Our goal all along was to make this not quite shrink-wrapped, but closer to shrink-wrapped than it ever was before. So the stuff just works. The things are ported to platforms and it works. And I think that's where the channel comes in because it gives [solution providers] that opportunity to sell what previously had only been a Big Boy sell. We've got [channel partners] who are doing very good on products like Itanium. One of the channel guys here sold a 1,024 Itanium node to Lawrence Livermore National Labs for a supercomputer. That's a channel product.

CRN: Looking at the emergence of, and the continued growth of, Linux, is that something that has led to incremental growth for Intel or are Intel-based systems that run Linux just displacing other Intel-based systems?

Otellini: No. Our data shows that for the most part, the Intel-Linux servers that are being deployed are displacing RISC-Unix servers.

CRN: So what's--looking ahead this year, maybe next year--what's Intel's biggest competitive threat?

Otellini: Probably the same one that we've had on the top of our list forever. And that's execution. Our requirement for our business model is to displace our entire product line every year. And we have to move to that next generation to take advantage of it and keep the business model moving. At the scale we're at now, that becomes increasingly challenging as you bring on three, 300mm, 90nm factories at once, and you go from zero to tens of millions of units on a new factory, on a new product line, so it has to be perfect. It has to be perfect.

CRN: So when you have a situation, like you had with Dothan [which was delayed in its launch to market], how much of an opportunity are you missing?

Otellini: I don't think we lost an opportunity for sales because we pushed it out a quarter. I think the opportunity was we could have started a little earlier and started utilizing that factory. But what we did instead was we said the yield on the product wasn't up to high-volume manufacturing standards. The functionality was very good. So we decided to use those wafer starts for Prescott instead. So we didn't waste any wafer starts and we were able to satisfy all of the Dothan demand with Banias. And now Dothan is yield-enhanced and we'll ramp it. So that was more of an asset-allocation decision than a product decision. If Banias had been further down in its life cycle and we needed to get a performance boost or something, it would have been different but it was still pretty young in its life cycle.

CRN: Intel recently said that it would provide 64-bit extensions for Xeon. When the Opteron was launched, AMD touted the backward compatibility factor. And the jury is still out as to whether those processors are being sold for backward-compatibility reasons.

Otellini: By definition, they are. No one is running 64-bit code. So it's all run in the 32-bit space, right? There's a Linux kernel out there that is 64-bit- enabled, but there's no applications to my knowledge unless they are home-grown. And Microsoft is just now moving out of beta mode. But you have to understand that we didn't just, a week before IDF, decide, 'Gee, let's put 64-bit extensions into Prescott and Nocona die.' This is obviously a multiyear effort, multiyear project. Our understanding of extensions and how to move, how to extend the Intel architecture from four, to eight, to 16, to 32 bits, to 64 bits is pretty good. You can imagine that we know what we're doing here and perhaps we might have some intellectual property in this area. Let's just say we're comfortable when the software is there to exploit it we'll be there.

CRN: When you're there and when you can talk about that level of scalability, will the market be there? And it's not just white-box companies, it's companies like Dell in that area [that] don't have a legacy there.

Otellini: It's a misnomer to call high-end enterprise computing, 64-bit computing. ... There's a lot more to it than the bitness of it. Even the extension architecture--neither ours or our friends' machines are full, 64-bit machines in the way Alpha was, or the way Itanium is. But there are 64-bit extensions and addressabilities. The attributes in enterprise computing are the classic RAS stuff. It's the ability to scale up. It's the ability to run those in a reliable fashion--five nines kinds of things--and really run the plethora of applications. I see the extensions being much more useful in the low end of the product line, the scale-out stuff, the clustering kinds of environments; applications that would take advantage of the memory addressability, so supercomputing things or scientific applications--but, even there, floating point matters. And so if it's a floating-point- intensive application, Itanium is going to win over anything out there--ours and our competitors'. But it really gets down to, 'What's the application?' We did a plot of PC applications, desktop applications, in terms of memory requirements over the history of the PC. And I can't remember the exact number, but it grows at like 100 megabytes a year, or some number like that. And if you look at where the preponderance of applications are today in terms of the need to address beyond 4 gigabytes, it's not going to be there until 2008, 2009. And, guess what, by the time you get the memory cheap enough to afford it, today that 4 gigabytes would cost you three times as much as the computer you're buying. So I think there's a natural curve here that as we watched the eight to 16, 16 to 32-bit transitions--the 32-bit transition took five years for the software to move after we moved the microprocessor. We're probably dealing with that kind of transition [now].

CRN: How important is the channel and solution providers to your business as you're looking forward?

Otellini: They are fundamental. They are fundamental to our business because they touch customers with the latest technology who otherwise would not be touched by [it]. And I also think there's a degree of excitement an entrepreneurialism that is still rampant in the channel that keeps us fresh in terms of the feedback they give us. It keeps us on our toes in terms of competition. And, probably it's good for the industry, healthy for the industry, otherwise this stuff really could be bought by anybody at Costco. And not everybody shops at Costco because they want that touch or they want that technology, they want that personalization. And as we look out, in time, even though we're trying to make the products simpler and simpler to use and install and configure as an industry, I still think it helps to have people who know the customers needs, their business, and know the applications. But it's fundamental.