InformationWeek

February 23, 1998

Andy Grove Interview
(continued)

INFORMATIONWEEK: I was there for the press briefing and your speech in the morning.

GROVE: You saw what I did at the Tandem/Compaq demo. That's a perfect example of that. Trace the progress of Tandem on the New York Stock Exchange or whatever it used to run on Tandem computers. The applications don't come a whole lot higher end than that. The whole specialty was high end, highly reliable nonstop computing and all that. And it was all proprietary. Then they moved it on one of the RISC processes, then Intel microprocessors became powerful enough so they started to move onto that. Then they are implementing in a clustered fashion so it's inherently scaleable, marrying their own cluster technology with Intel- based computing tec hnology and the process just continues on and on. It comes around by the marrying of their own added value with our processor technology.

INFORMATIONWEEK: I wanted to ask you about the thin- client market. I guess Intel really is not a player in that area. A few months ago you were saying [you would] consider using [older] Pentium processors for that market. Have you changed your approach on that yet?

GROVE: I am put off by your saying that we are not a player in that [market]. Thin client -- I don't know what you mean by it. The Oracle NT ...

INFORMATIONWEEK: That's a consumer device.

GROVE: No, it isn't. They have ambitions of getting the set-top business. But NCI was formed to be basically a database query computer. When you cut through a lot of the rhetoric that's what it is. I think it's based on both ends on Intel architecture.

INFORMATIONWEEK: I think people generally accept that most of the people who have shipped reasonable numbers of these thin-client or network computer devices are not using Intel processes. Maybe NCI is, but ... it's just that people consider NCI not to be shipping a lot of product yet.

GROVE: No, they aren't. Neither is anybody else. I don't mind you contradicting me. I'm puzzled. Who is shipping anything of this?

INFORMATIONWEEK: Some people are shipping tens of thousands, possibly hundreds of thousands.

GROVE: Of what? ... Let me back up for a moment without us arguing. I don't think this category has gone anywhere yet. It hasn't gone anywhere yet not because of problems of designing the client or microprocessor or otherwise, but because it's a systems issue and it takes quite a while to make it a viable proposition. As much as it becomes a viable proposition, we are working with every manufacturer, network computer lean-client type of device, whether they are using BUS architecture or another. We take this as a very interesting proposition whose impact is goi ng to be replacing a generation of dumb terminals in which we did not participate. This is a market territory that we have not been in where with the turn of the screw from dumb terminals, we go to a lean-client paradigm. The goodness of Intel architecture and the commonality on the device on the server and client side has value in signifying the operation of the total system. We are very devoted to that.

INFORMATIONWEEK: Will you be using the [Covington] on any sort of thin client?

GROVE: I don't know of any such work at the moment. We could.

INFORMATIONWEEK: Any Pentium II-class chips for that market within the next year?

GROVE: Not in the ones we're working on. On the other hand, it's all a matter of time. This category has been in talking stages for a couple of years and it hasn't gotten very far. If the rate of adaptation stretches out another two years, that's going to be the only thing we'll manufacture, so that's what we'll end up using.

INFORMATIONWEEK: What about the [low-priced desktop PC], corporate desktop arena? Where do you expect that to go in the near future in terms of price?

GROVE: It's got to be in the thousand something dollar -- $2,000 price range. Get more and more value for [that price] like you have every year ... heaven knows how many years.

INFORMATIONWEEK: So you don't see average selling prices for the corporate market really declining?

GROVE: I really don't because the compromise we have to make represents sacrifices in longevity which represent human work. IT departments understand that. Just using an IT organization in that, they understand that every time they have to go out and upgrade a machine it's costing them far more than buying a machine that is likely to take more time before it gets saturated with applications.

INFORMATIONWEEK: I get the impression from the announcement earlier this week that you made with IBM and the management initi atives that -- and I hope this is true -- in some cases, Intel plans to have more of a direct relationship with the IT customer for certain technologies. If that's the case, what is the advantage of having that kind of direct relationship? Does Intel have a better sense of being able to influence technological changes in the enterprise that way?

GROVE: I'm OK with both the beginning and the end of your question. I'm not OK with the connection. What we are doing with IBM is combining our technologies to allow corporate networks to be more easily managed and cost- reduced. [That] is completely consistent with that drive toward the network-rich computing that I described earlier.

We have for some years realized that we have a contribution to make to a corporate IT department because [of what] our planning horizon and our technological horizon is. The planing horizon of an IT department, particularly in larger organizations, is very similar to our product-planning horizon. In that period of time, it is very relevant to know what the fundamental platform technology is going to do -- microprocessor, BUS structure, connectivity structures, and the like. We've found that whenever we went to talk to IT departments, we had very receptive ears because these people were making the decisions, not for the next year, not for the next purchase cycle, but an infrastructure decision that was meant to last multiple numbers of years.

So, we have put out a non-commissioned field force, a field force that does not sell anything, but on a strictly consultant or advisory capacity worldwide several years ago who have been taking our message out to IT organizations, bringing factory expertise where it is appropriate, and we continue to do that.

INFORMATIONWEEK: How large is that consulting sales force?

GROVE: Several hundred people worldwide. I don't remember exactly.

INFORMATIONWEEK: Do they tend to work with original equipment manufacturers?

GROVE: Absolutely. Whenever the OEM wants to work with us. We don't force them. We do not make computer equipment recommendations. We are OEM neutral. When it comes to technology, many of our OEMs rely on Intel technology so our technology stores tend to be shared among various OEMs. We cooperate with whatever OEM wants to cooperate with us, but our story is technological and not product-oriented.

INFORMATIONWEEK: How do you divide responsibilities between you and Craig [Barrett]?

GROVE: Well, for starters the division is not that precise. We are in many instances a plug-in replacement for each other and sometimes, since we work pretty closely with each other, I can pick up the pieces if he's out of town and he can pick up the pieces when I'm out of town. It gives us a lot more leeway. For instance, I'm not going to be present at the next executive staff meeting and I'm perfectly OK with that because Craig is going to be here. So you're not going to find a precise division that way , but if you draw two bell-shaped curves which overlap in the middle, he is basically operational and I am dealing more with larger high-level strategic issues and external contacts. But again, it's not exclusive because he does a lot of that also and, like it or not, I poke around in operational issues every once in a while, too.

INFORMATIONWEEK: Do you see that changing in the future, your role?

GROVE: Well, I see him getting more involved with strategic issues and getting more involved with external issues and that is what is supposed to happen. We are at different stages of our career curves and he needs to move in that direction ... and he is.

INFORMATIONWEEK: Analysts we talk to say that one of the key questions about Intel long term is who will succeed you. Your name is very strongly linked with Intel the same way that Gates' is linked with Microsoft. Can you describe a succession plan?

GROVE: Well, I don't think that's particularly appropri ate.

INFORMATIONWEEK: Another way to come at this is to ask what kind of company will Intel be in five years -- especially vis-à-vis the enterprise customer?

GROVE: My hope is that we are going to be even more relevant technologically than we are today. Remember what you asked earlier about the technological roadmap? I think the technological roadmap is richer and more relevant to an enterprise customer today than it was five years ago, and I think the trend is going to continue and five years from now we are going to have more things to say and more things to describe that are of interest than today. We are broadening our participation, earlier what you called -- your words, not mine -- the "nontraditional investments" kind of thing. That gets us into visualization, into networking, into network management, into how to describe data-mining efforts that was an Intel laboratory effort. All of these are bits and pieces of technology that are relevant to an enterprise custom er and there are going to be more of them five years from now.

INFORMATIONWEEK: What might some of them be?

GROVE: I really don't want to suggest that I have a big surprise up my sleeve, because we don't operate that way. All of the things you are thinking about are networks, network management, visualization, transaction processing spreading upward and downward exactly how we describe it, and of course -- again, it's not a card up my sleeve -- but we're going to usher in major new architecture in the IA 64 category that is going to be absolutely relevant to the enterprise user. So, five years from now that's going to be well under way and a work horse, I would hope.

In terms of corporate characteristics, my hope is very simple. I hope Intel is going to be no different than it is today. That's not to be taken for granted. Left unattended, a corporation tends to age and slow down. I think Craig's contribution has been enormous in continuing the vitality of Intel and the p ulse rate and the exacting adherence to [a] value set, and I trust it will be at least as good five years from now as it is today.

INFORMATIONWEEK: Could you address the cost of ownership [issue]? Do you see much progress in the future in the cost of ownership?

GROVE: Including the IBM announcement that we described, I think we are making pretty major progress in that. I think the portion of client [technology is] approaching 50% -- again this is not an exact number, but an impression I have of a number I have seen. Management is getting it on the server side. The interoperability between ourselves and other companies, this is how cost of ownership is going to go down.

INFORMATIONWEEK: Thank you very much.