'WE PUSH TECHNOLOGY AS FAST AS WE CAN'
Few technology executives have had the success of Intel Corp. CEO Andrew S. Grove. In one of the fastest-changing businesses today, Grove has managed to keep Intel ahead of the pack in everything from PCs to videoconferencing. Along the way, Intel has captured some 90% of the PC microprocessor market and, by charging top dollar for its chips, become one of the most profitable companies in the world. Correspondent Andy Reinhardt spent a couple of hours with Grove at the company's headquarters talking about Intel's latest quick change--engineering chips specifically for cheap PCs and low-cost devices, a first in Intel's 29-year history.
Note: This is an online-only, expanded version of the interview that appears in the December 22, 1997, issue of Business Week.
Q: Intel has always produced high-performance chips and, over time, moved the technology downstream. Why change that now and make new chips for cheap PCs and other inexpensive digital devices?
A: When all is said and done, we are a manufacturer--a high-volume manufacturer. And all the investment we put in technology and manufacturing capability is meant to produce things by the hundreds of millions. So for us to walk away from a market whose size is going to be measured in tens of millions of units per year, maybe bigger, is inconceivable.
Q: Why not just use older technology, say, Pentium MMX chips, for low-cost devices?
A: [Grimacing] You can see my reaction. We are what we are because we push technology as fast as we can. Our whole belief is that technology is good, and more technology is better. How could we slow down technology? It's not good for anybody: for the software developer, for us, and most important, it's not good for the consumer.
Q: Is this a strategic inflection point for Intel? A "10x change," as you call it in your most recent book?
A: I don't think so. I don't think we are dealing with any technological phenomenon. What Intel has done in 1997 is to introduce two [chip] generations in one year. So, you've got doubled flow of technology and products down the hierarchy. The product that we are in the throes of obsoleting today, we introduced in the early part of '97. So the computers that people can buy for a low cost are particularly good. Before, this stuff was warmed-over inventory liquidation--already obsolete when it hit the shelves. These machines are not. It could be a major strategic error if we didn't respond to [the rise of cheap PCs]. Over time, if we allowed it, it would grow into a competitive threat. We won't allow it.
Q: What would be the nature of that threat?
A: If we were slow to respond, we could create a vacuum for other people to get established in that portion of the market. But even in the absence of competition--in the absence of price pressure--we would be doing something very similar to what we are doing today. Because we didn't develop Pentium II and build all those factories to use 15% of the volume for Pentium II. We built them to fill them 100% with Pentium II.
Q: Are you more concerned about Intel-compatible products that could gain a foothold, or different approaches that are non-Intel-compatible?
A: You know what I'm most concerned about is our own products getting stuck at the low end, and the flow plugging up. I really don't mind the change in the market so much, as long as the top-to-bottom technology flow continues. If the flow stopped or slowed down, I would mind it a lot more.
Q: You recently pointed out that personal computers using the Pentium II chip had dipped under $2,000 in record time. For consumers, that's good news. But for Intel, doesn't this mark the shortest time yet for charging a premium on your latest chip?
A: Well, wait a minute. The volume is much higher. How it will exactly work out, I don't know. The notion that we sit around and say: 'Let's slow it down so we can charge more money...' I mean, this is the room where these discussions take place, and these walls have never heard that. That's not instinctively right, and our instincts haven't been bad. When you look at once proud-and-mighty corporations that have fallen on hard times, they've always fallen on hard times by optimizing [for the] near term vs. a belief in the long-term efficacy of introducing new technology.
Q: Does the reorganization you just announced signal that you see segmentation as being even more of a change than you thought it was previously?
A: No. We don't reorganize from one day to the next. This has been coming for the last couple of months. We need to line up our efforts one-on-one with the segments. It's no more, no less.
Q: This suggests that fundamentally your new strategy, if that's not too strong a term for it...
A: It is too strong. It's an acceleration of our strategy to supply the low end and a decision to supply it with products that are designed for that target, rather than cascading down. We're designing products for servers, for workstations, for basic computers. We are consciously going after the low-end part with engineering.
Q: So is this a fundamental shift in your business model? In the next few years, are you going to turn into a producer of megahigh volumes of $20 processors?
A: No. No. No. No. No. No. No. I don't think so.
Q: I guess you mean no?
A: You're trying to get black-and-white out of a very gray deal here. First of all, there are $20 processors in our portfolio today, in the embedded area [chips that go into other products such as cars and printers]. I don't think the news is going to be in $20 processors. I think the news is going to be in penetration into a wider population--the proverbial 60% of U.S. homes that don't have a computer. Why don't those 60% have computers? What we're trying to do is put a need in there, and at the same time, make this stuff affordable. And if that formula starts working, we want to make the same kind of money we're accustomed to. And the only way we can do that is to design for the target.
Q: Is segmentation and addressing low-end markets going to slow your earnings growth?
A: I don't really know. But whatever it's going to do, we are going to be motivated by participating in each of these segments to the fullest extent of our technical and marketing capabilities.
Q: Analysts are projecting falling average selling prices for Intel next year. Does that worry you?
A: What worries me is being in the forefront of the changes that we are embarking on. If we don't let any of those changes leave us behind, we're going to get an appropriate mix of business: high ASPs [average selling prices] and low ASPs and growing volumes. Losing nothing in the process, as the market shifts -- then I'm happy.
Q: Walk me through how your moves into workstations and servers will pay for your move into the low end.
A: I can only do that in the most superficial fashion, because first of all, I don't know. What is very important for us, and a driving force, is to participate at both ends of the wire, and participate in a leading capacity at both ends of the wire. And I think the formula is going to work out.
Q: You've said publicly that over time, you see Intel's gross profit margins slipping from 60% to 50%. Have you identified any kind of time frame?
A: No. And we are not prepared to, either. I don't want to be wedded to a particular percentage. How long is it going be? We'll make it as long as we are able to.
Q: There's a lot of talk about new markets emerging for tv set-top boxes, Internet phones, and pc/tv setups. What product category excites you the most? Which one offers Intel the most opportunity?
A: These are different questions with different answers. The one I'm most excited about is streamlining the business computer to a network use. I've [long] felt that the predominant use of a business computer is as a communication device, and doing that right--infusing networking and management, a proper balancing of servers, intranets, and all that stuff--to me is the most fascinating. From a business standpoint, and from a technology standpoint, what I find more fascinating is the entertainment computer in the home. The stuff that has videophones and digital imaging and very alluring browsers, with lots of technology to move the interface to another level of capabilities. Actually, it's hard for me to choose between those two.
Q: The client device that people need to access the Internet infrastructure doesn't necessarily have to be an x86 computer...
A: It doesn't, but it will. You would have to have a competitive advantage to break compatibility with the installed base. How are we going to allow anybody to have a competitive advantage over what we are providing?
Q: What about the cost?
A: It has to be compatible, in my very highly prejudiced opinion, and it has to be very high performance--provided you don't price it out of the market. The fact that we are pouring the technology down into mass-affordable devices may turn out to be a long-term advantage, because you don't have to make compromises that will make that interface less friendly. The other aspect, if you're looking from the Intel standpoint, is that every time one of these devices gets shipped, one-tenth of a server gets shipped someplace that you never see. We intend to make that a major part of our business. I think there's going to be a huge growth in that, and it's not going to be particularly visible.
Q: What if instead of making processors even faster, you built integrated chips that had more than just a processor on them?
A: That says, no new software is going to come out that requires more performance, and therefore, all we're going to do is reduce the existing device to a lower and lower cost. I don't believe that. That's a very fundamental point of departure.
Q: But you gain other advantages from integrating on-chip...
A: Is the computer a telephone? Has it reached that state? That is the existential question you are raising. I don't think so.
Q: I guess that's the existential answer you are providing...
A: And I'm providing it absolutely without the slightest hesitation!
Q: I still don't understand why you might not want to build some kind of integrated chip on a non-leading-edge production line.
A: We won't. Because CPU performance, in our opinion, continues to be a critical determinant of the efficacy of the computing device that we are going to load down with state-of-the-art browsers and communication devices. We don't want to hobble the computer.
Q: What are some of things that are going to push us in the next year beyond the plateau that we've been on from a software standpoint...
A: We have not been on a plateau. I disagree with that. You load the 90-MHz Pentium, which is a two-year-old product, with the kind of software that a common PC needs to use, which is a modern browser--Navigator 4.0 or Explorer 4.0--Office 97, a few communications packages, and the computer won't even wake up. The Internet is the killer application that people have been talking about. Internet applications are moving very rapidly...you wouldn't want the browser that you started with two years ago any more than you would want your 90-MHz Pentium.
Q: As Intel sells chips into new markets, such as set-top boxes, your interests may diverge from your longtime partner, Microsoft. An analyst commenting about the handheld market recently said you're "sleeping in separate beds." Is that true?
A: I don't think we've slept in the same bed, ever.
Q: But what about the children [products] you've made together?
A: Well, that's right. But you can do that in separate beds. You can do it in the back of a car. (Laughs.) I dare you to use that one!
Return to main story