CEO Roundtable: Bright spots

This year our CEO Roundtable focused on the many changes going on in the electronics industry, how communications no longer reigns supreme, and that maybe—just maybe—there are some brighter times ahead.

Staff -- Movers & Shakers, 11/1/2004

Participants (Clockwise from left): Aart de Geus (Synopsys), Keith Jackson (ON Semiconductor), Wim Roelandts (Xilinx), Bob Swanson (Linear Technology), Alex Lidow (International Rectifier), JamesTruchard (National Instruments), Ned Barnholt (Agilent Technologies), Wally Rhines (MentorGraphics), Jerry Fishman (Analog Devices), Michael Marks (Flextronics).

Q:We have had a tough three or four years in the electronics industry, although some of you have seen some changes in the last three to six quarters. I don't want to dwell too much on the past, but what have you done to try to keep your company strong and viable and to be here tonight as an example of what's been going right in the electronics industry?

Michael Marks (Flextronics): Well, I'm sure I have the worst stories. I'm happy to start. We closed 100 backrooms and wrote off $2.5 billion. Can anybody top that here?

It worked. We're back making money. Next quarter, actually, will be a record-profit quarter for the company. It's still not enough, but I'm sure we did what everybody else has done here. We just whacked away. We whacked the cost, consolidated, and we cut out everything we were doing that wasn't necessary. And now we're back running a decent business again. It was not fun, but it's much better today.

Q:Is it sustainable going forward? Do you see any possibility of it leveling off or even declining?

Michael Marks (Flextronics): I was on Wall Street last week, and the Wall Street guys are completely panicked about any little uptick in inventory and totally focused on small things. Sure, it's going to cycle. But my point of view is that 2001 is not happening again now, and it won't happen again in our lifetime. You only get a huge crash like that when there's no caution anywhere. Today, and I'm sure you guys will all agree with me, there's an abundance of caution at every layer: suppliers, assemblers and OEMs, retailers, and consumers. As long as there's an abundance of caution, you're not going to have something that just explodes and we all go back into batten-down-the-hatches mode. I think that's gone for a long time. But that doesn't mean business can't cycle a bit. It certainly can.

Ned Barnholt (Agilent Technologies): Today I agree everybody is just being very cautious. They don't want to be caught with too much inventory, they don't want to be caught with too much capacity, and they don't want to be caught with too many people, which is one of the reasons why you're not seeing hiring resume at the levels that we saw back in the latter part of the '90s and 2000.

At the same time, there's been a lot of structural shifts. More and more of our markets have moved. You look at contract manufacturers, even semiconductor assembly and test are more in Asia today. Four years ago, 25 percent of our business was in Asia. Now it's 45 percent. Part of it is to support the customers that have moved there. Part of it is to support indigenous companies that are growing there. That's a big shift in the global economy in electronics. I think there's a caution, though, that makes everybody want to make sure they don't get caught. We all have long memories, so we don't want to get caught like we did in 2001.

James Truchard (National Instruments): Our response was to expand our R&D. We've doubled our R&D staff over the last four years, and as a consequence, we've been able to introduce a lot of new products over the last year or two. So our approach was to double down and make sure, as we came out of the downturn, we would have a very strong position.

Q:But you had a pretty good position in the middle of the downturn too. Would you have made the same investments if you would have suffered the way the semiconductor industry did?

James Truchard (National Instruments): Probably not. We were fortunate enough to be in a position that we could invest. It was good because not many people were hiring, and we could hire top graduates. We picked up a few seasoned folks to give us a little direction with all the young talent.

Q:This industry was started to a large extent by very bright people, but also by a lot of people with a real sense of bravado. It sounds like bravado has disappeared out of this group. What happens next? Are you more vulnerable as a result of that?

Keith Jackson (ON Semiconductor): I think the key for us has been to get intense focus on what's important to the customers. At least from our perspective, there's really only three things they care about: products, cost, and service. So if it's not in the category that's going to provide something they want and need now, we've just had an intense focus and shed that piece of the operation. Get the cost off the books, and just really focus on those three elements.

With that, I think you gain a share of independence from what's going on with the macro environment, because we've got very little control over that macro environment. What we can do though, is make sure that the company stays lean and stays focused on those elements.

Q:Isn't there a threat though, that you might lose out on opportunities that are going to appear three, four, or five years down the road?

Keith Jackson (ON Semiconductor):

Again, I think if you really focus on what the customers want, you're probably not going to be very vulnerable to that because you're going to be spending enough time with them to know what that is. Now, none of us is going to call it 100 percent right. But you shouldn't have major surprises if you focus on the customer.

Aart de Geus (Synopsys): It seems to me that the term "downturn" is not correct. There may be a downturn overlaying things, because there are downturns every year and a half. But this was a restructuring and a move to a next-generation phase of the industry, for a couple of reasons.

One, if you look at the end products that are starting to dominate, all of them are consumer products. It's computation plus communication all packed in one. Two, consumer products—very high volume and very low price. Three, a dramatic change toward globalization, both in potential new emerging end markets and markets that are of a lower socioeconomic level and therefore, again, cost sensitive. And four, many more new employment markets also have lower costs. So the common denominator here is cost, cost, cost, and cost.

The numbers may not be exact, but you could say fundamentally in the last five years there have been 200 million consumers coming online in China, 200 million in India, 200 million plus in Eastern Europe. That's as much as the rest of the western world used to have. Granted, they spend less, but they're going to spend, and many are very well educated.

Wally Rhines (Mentor Graphics): Our customers, in the business that Aart [de Geus] and I are in, semiconductor companies and system companies, have nearly doubled their unit volume since the 1999/2000 boom, but they haven't raised their revenue anywhere near that amount. In fact, it's pretty much where it was at the peak. So as a result, they simply have to do more for less.

In our case, the way we can take advantage of that need is to focus on things that impact their cost in a big way. Our industry didn't typically have a major focus on manufacturing, on the software for lithography, resolution-enhancement software, on the design-for-test software—things where you could point to it and say, "This will save you hundreds of millions of dollars." By focusing on those things, those are growing markets in spite of the tremendous pressure our customers are under to get their R&D down to a reasonable percent of revenue. There are some things they'll spend money on, and that's what has been successful.

Q: The things you've mentioned, aside from manufacturing, are very important in bringing the cost down, but I'd like to ask everybody, how do you reset the consumer mindset that stuff should be free? I just don't think you should be able to buy a DVD player for 39 bucks. You just shouldn't be able to get it. Wally Rhines (Mentor Graphics): I don't think you do. You do not have the right to reset that expectation. The market has the right to do that. So you have to live within that, and we simply have to find ways to make money in that scenario.

Michael Marks (Flextronics): Why shouldn't you be able to buy a DVD player for $39?

Q:The cost of manufacturing is higher than that.

Michael Marks (Flextronics): No, it isn't.

Q:The only reason it's not higher than that is because they're using mechanisms and pickups that have been in inventory and basically written off. I mean if you were to go buy a state-of-the-art drive mechanism, there's no way you can build a $39 DVD playe r. If you used some inventory that was written off two years ago, you can build it.

Michael Marks (Flextronics): I don't think you're right.

James Truchard (National Instruments): When you look at your cost, and then somebody's selling this $39 DVD player, it is a concern.

Q:As the market starts converging into a lot of different areas, it seems like the one theme that comes out over and over again is the consumer market. Is that an opportunity or a threat for all of you?

Ned Barnholt (Agilent Technologies): I think it's the inevitable future of the industry. If you go back and look at 20 years ago, the industry was much more dominated by infrastructure and large IT- or communications-related systems. But electronics has migrated into the home, into the consumer markets. So the bigger value-added today is in the consumer space. That trend is going to continue as more of Asia and Eastern Europe and other parts of the world come into the world economy and their capacity to buy increases.

So the question is what the implications are for us. I think the implications are, as a more consumer-related industry, we are going to have a lot of cost pressures. We're probably going to be more volatile. Seasonality is going to be different. We're going to see seasonality much more like consumer products. The cell phone is now a consumer product. It's not a communications product. So these things have changed the way we have to manage our businesses.

James Truchard (National Instruments): Technologies like USB and FireWire and PCI Express have become consumer technologies. And as consumer technologies, they've come to a very low cost point. For example, take PCI Express. If you look forward, you'll be able to build a server farm just using consumer PCs. When you can do that, that changes everything for higher-end systems. So this technology, because of its volume, just drives down the cost of everybody's products.

Bob Swanson (Linear Technology): Five years ago, if you mentioned consumer products to my company, you'd have to wash your mouth out with soap. But the fact of the matter is that the high-end digital-consumer products like these full-featured cell phones, iPods, and digital cameras—I'd hate to say this in front of my competitors—but they saved our butt two years ago when the market was starting to come back up. These consumer products insisted on a very small form factor and a lot of complex analog. And the first company that could put that in a small form factor and do it in a cost-effective way won a whole lot of business. So we're in a consumer business like we've never been before, and it is volatile.

In fact, unit volumes are doubled, just like you said, but it's not because we're selling the same products at half the price. Our ASPs [average selling prices] are down, but our gross margins have never been higher. It's a totally different mix going on now. All these consumer products, it's small, small, small, and the good news for us is they need some really good analog stuff. That's a constant. So we love the consumer business now.

Q:But you analog guys have an unfair advantage: You don't have to go and push to the next process technology.

Bob Swanson (Linear Technology): We never go to a higher process technology to lower the cost, because that's not the way our model is. It's about building in functional value for the customer, and what it costs us is neither here nor there. But in order to squeeze a lot of this high-end analog into a smaller and smaller area, we're forced to go to the next level in process technology, but it's what the digital guys had capped off years ago.

Q:It's still nothing like your buddy from Xilinx down the table is facing.

Bob Swanson (Linear Technology): A totally different world.

Wim Roelandts (Xilinx): It is a different world, but the industry has changed. People maybe haven't realized that, but we are in a different industry, where people are much more critical when they buy technology. You don't buy technology for technology's sake anymore. You buy technology because you have a need for it. That couples with the overcapacity in the industry, because with all the downsizing none of the big companies have disappeared.

Jerry Fishman (Analog Devices): I think one of the constants that will continue is to try hard to stay away from commoditization. Yeah, there's going to be $39 DVD players, and there's going to be a lot of people that chase that same $1 component that goes into a DVD player. But I still believe, and I think listening to Bob [Swanson] probably he does too, that there's still an opportunity for huge and massive innovation out there. I think those companies that are in the mode of just forcing the cost down and commoditizing everything, it's hard to make a lot of money doing that. There's always somebody who's willing to do it for less—whether or not he makes money.

James Truchard (National Instruments): That's exactly right. So you've got a $1,000, 8-megapixel camera using the latest parts. Then you've got the low-end one. And the high end is going to have a good margin and the low end won't.

Jerry Fishman (Analog Devices): It's really the cost per function that you try to narrow down on, as compared to just absolute cost. You try to provide more and more functionality. I think even Intel, if you listen to some of the dialogue, lately is saying that just promoting higher-speed microprocessors doesn't pay anymore. If you can do something for the customer that's different than what everybody else can do, there's plenty of money to be made. That's where you ought to reside.

Q: If you can run your business on that leading edge, and you can always be in the leading-edge product and then out of that market before it becomes a commodity, that's great. But whoever is supplying the part when it does become a commodity is basically going to go out of business.

Jerry Fishman (Analog Devices): I don't think so. I think you're the one with the volume. You've got your cost down to produce that function. More digital cameras sold than film cameras this year. There's a lot of commoditization in that business, and yet, if you can produce a product that enables a little better picture, a little better image, or a little more processing of that, you can still get a premium. I think the real key is to stay away from the really dark and dirty low-end, where nobody is willing to pay for performance. Even when the higher performance begins to commoditize, your costs are lower than anybody else because you made more than anybody else.

Q:But that doesn't quite translate over to the digital chip guys. None of them are here tonight.

Bob Swanson (Linear Technology): That's right. And again, I think the analog guys, we've got one of maybe the two businesses left where a company can still create a product and differentiate from what the pack is doing. So that's what we have going for us. I think we take credit for knowing that 20 years ago.

Jerry Fishman (Analog Devices): A very famous CEO in the analog business has been quoted widely as saying to all his guys, "Don't tell me where the competition is—tell me where they're not." That's the key.

Q: In this kind of market, where stuff is getting commoditized very quickly and you have to stay ahead of the herd, doesn't that make innovation in R&D even more important? And yet, at the same time, you're having to manage the cost equation. How do you ba lance those two things?

Bob Swanson (Linear Technology): You manage the innovation part first. Then you look to see whether your bottom-up cost allows you to sell this product. If you do the innovation part right—if you solve the hard problems for customers at a price that they have no alternative to do it cheaper, they will buy your product.

Q:Your customers are going to be the people that the OEMs and so forth can sell to. Consumers will be going through two levels. So you have to sell them and convince them on the one hand that there's an end user, a consumer need for them, right?

Bob Swanson (Linear Technology): The answer is yes if you're talking about cell phones. I used to smile at our conference calls with investors who would say, "You're not in the handset business at Nokia or Ericsson, so you must not be in the cell-phone business." Well, I think all my competitors know that we sell an awful lot of product to other suppliers, and Samsung is a good example. Samsung has taken over the full-featured cell-phone business. They're the ones that put in really good cameras and other features. So all of a sudden Wall Street has realized that there are people in the cell-phone business making money besides Nokia and Ericsson because Samsung focused not on vanilla cell phones, but on cell phones that had a difference.

Ned Barnholt (Agilent Technologies): I think at the end of the day there's a real strategic question for a lot of companies. Do you focus on just the commodity-related businesses or do you really focus on innovation?

The strength of this country, in many ways, and the strength of many of the companies in this room, is innovation. So you almost have to go back to your roots and say, how can you differentiate? And if there is going to be a $39 DVD player, and we're asked to provide components into that, maybe that's not a business that we should be in.

For example, we provided a lot of the chipsets for HP printers over the years. As printer prices have collapsed, we've walked away from the business. It's just not a good business anymore. It's a commodity business. There's no margin in it. So we focused on other things. That's a strategic choice that a lot of companies are going to have to make. Do you compete on the basis of commodity, in which case you have to be world-class? Or do you compete on innovation? I think for most of us, innovation still has to be a key part of what we do.

James Truchard (National Instruments): One of the areas that helps is system integration—taking a lot of components and putting them onto one chip. And there the key is design tools, which our colleagues here tonight make. I see companies who are trying to do that without the right tools, and they're running into trouble. They can't turn the chip fast enough to get to market fast enough, and by the time they get to market, it's too late.

Alex Lidow (International Rectifier): Most technology companies define themselves in terms of their technologies. That's their core competence. They tend to look out from their core competencies. About four or five years ago, as a way to deal with volatility in the market, we selected a few core challenges that we felt would be around for many, many years.

One of the core challenges is that we knew that microprocessors and other high-performance chips would become thermally limited. That in itself generates an enormous number of challenges for power management, which is all we do. So I can't help Intel with their chips, but I can help deliver the power.

What we discovered was, if you just extrapolate using even Intel roadmaps, you'll find that no silver bullet—from a power-management point of view—will solve that problem forever. What that did was it created a mosaic of technologies that's constantly changing with time.

Q: As we start getting into more converged business models, as we start getting into more converged markets with consumers, does that make it harder to differentiate your company? And does it mean that your competition becomes far more intense?

Bob Swanson (Linear Technology): I think the Intel case is a good example. They look out five to eight years, and so now everybody knows what the finish line is. So for any new technology that's a business for us, unless we can figure out how to do something nobody else can do, Intel has kind of screwed it up for us by drawing a finish line. If the finish line stays constant, that's not a place for us to be. We need to be in a place where we keep pushing the finish line further ahead, and that satisfies certain customers. What's going to happen here is that everybody knows exactly what Intel wants. Two or three suppliers is plenty. There's 50 guys in the analog business. That means a whole bunch of them aren't going to be able to make a buck.

Jerry Fishman (Analog Devices): In some ways, that applies to most standardized businesses, right?

Bob Swanson (Linear Technology): There's a standard, and if the standard is good enough there's no need for anything better.

Jerry Fishman (Analog Devices): There's such a large part of the semiconductor market where the units keep growing and the dollars keep shrinking and the market stays flat. If you take a look at modems, I don't think there's anybody who's ever made any money selling modems. Any kind of modem—56K modems, DSL modems, cable modems, 802.11 modems.

When you're in these standard-based businesses, as Bob said, you might get the order if your performance is a little bit better, but you don't get any more money if your performance is better. Those turn out to be lousy businesses.

James Truchard (National Instruments): So the key then is to use that standardized roadmap that Intel is creating and see what peripheral businesses develop. The key is to understand how PCI Express or whatever is going to change the marketplace and find opportunities that are adjacent to this commodity business.

Bob Swanson (Linear Technology): I agree. In the case of Intel, they have these specifications for core power, and everybody knows what they are in advance. But there's a lot of system power opportunities. So you're right. It's sort of like saying that if you ride to the sound of the guns, you're going to have a lot of company.

James Truchard (National Instruments): For example, in test and measurement, PCI Express gives us an order of magnitude improvement over previous standards, and if we figure out how that's an advantage, we use it.

Michael Marks (Flextronics): I would like to jump in here because this is kind of a weird group in the sense that this is a semiconductor-driven group. So, I would just like to take a moment to back off and take a look at the world not as semiconductor people see things. And I'd make a couple of comments.

Commoditization is not a bad thing. I'm going to strike a blow here for commoditization. I think $39 DVD players are fine. Commoditization is only bad for everybody who's not a believer. I mean, PCs are a horrible business unless you're Dell, and cellular is a horrible business unless you're Samsung or Nokia. You'd better be No. 1 or 2 or don't bother. But commodity businesses are always very good for the guys who are No. 1 or 2.

So, first, I want you to acknowledge that. If you're not going to be 1 or 2, you've got to go and do something different. You don't want to be No. 5 in a commodity market, then it's the worst business on the planet.

The second point I'd make is that, if you stand back and look at the supply chain, if you look at all the players who are involved—from the raw materials to the components to assemblers to OEMs to retailers—what's happening, despite the whimpering of some of the people here, is that only the guys who make the components and the guys who take the product to market are making money. That's a fact. It's probably not an accident that there's no OEMs sitting at the table today.

If you look at the OEM business, the OEM business is lousy today. In fact, OEMs are shrinking in their value to the world because there are component companies, like my associates sitting here, and assemblers like us going directly to Best Buy, Amazon, eBay—and so traditional retailers are losing out, traditional OEMs are losing out, traditional assemblers are losing out. And actually of the only two groups that can win, one of the groups is you guys—the component manufacturers. You can have all the discussion about where you want to play and how you want to play, and you have to be good and all that kind of stuff. But in fact the margin in the chain largely resides with the companies that are here today. So you're in the best part of the business.

I'm actually not in the best part of the business. I'm in the group that's squeezed, because assembling products today, they're all more or less the same—which is why we're getting into designing components and other kinds of things. We're doing a lot of things in our company to get out of that. But it's actually a very interesting question that doesn't get talked about much.

What's the future of the OEMs? It's not very good, I think. They're actually the ones who will get squeezed by commoditization, not necessarily the component guys.

Bob Swanson (Linear Technology): I think we all can get squeezed if we allow it.

Michael Marks (Flextronics): Absolutely. But if you look at gross margins among these groups—gross margins on the raw materials, gross margins at the components, gross margins at the assemblers, gross margins to the OEMs, gross margins at the retailers—the best margins and the most sustainable margins have been at the component suppliers. If you're a lousy component supplier, tough luck.

Q:But even the component suppliers have a different landscape, because you really don't sell directly to the system designer. I mean, maybe you have to be in a reference design. It's a very different thing than a one-to-one relationship like you've had in the past. I think the component suppliers are having to share their revenue with multiple levels of the food chain even before it gets to the so-called OEM.

Wally Rhines (Mentor Graphics): The thing that ties together what has been said, is that it's great if you're No. 1. The point of being the leader in a new emerging market is that it's something you can be No. 1 at. So, when you're No. 1, whether it's a commodity market or an emerging market, it's a good position. The industry wastes too much money with the No. 4 and No. 5 and so on because the market doesn't need them. All they do is press the margin out of product sooner. All of us, and the OEMs as well, ought to be spending more time thinking about doing things that other people aren't doing—things that they can be No. 1 at. Build the strategy around that if you want to make money and you want to have growing markets.

Bob Swanson (Linear Technology): That's exactly right. The business-school rule says that if you have somebody who's making 30 percent to 40 percent after-tax margins, it will attract a crowd, and pretty soon, there go the 40 percent margins. And that's exactly what happens in this quote-unquote commodity market. Because other people can also do it. You can't defend your position forever.

Q:But doesn't every business become a commodity business at a certain point?

Bob Swanson (Linear Technology): Some portion of your business can become commodity. So you hang on to it as long as you can, and then you move on. But you'd better be fast on your feet and you'd better be innovating. You'd better be putting out a steady stream of new products, or else all your businesses will be commodity.

Q:The price of producing ASICs has gotten so off the charts that we're no longer in a commodity business, we're actually being hemmed in by the cost of developing chips. How does that impact everybody at this table?

Wally Rhines (Mentor Graphics): I think it says we're going to see another stage of customization. The process of designing 100-million-gate chips, tooling them, and putting them into manufacturing has gotten to a level where only market segments with tremendous volume can afford it. It just means that we're going to figure out other ways. Very likely, we're at a turning point. Whenever you get to a stage where things become impossible, and you're not seeing further improvement in cost and capability, then new architecture innovations and new approaches appear.

Q:How do you differentiate between an ASIC and an application specific standard product [ASSP], not from the point of view of how it's sold, but from the point of view of how it's accounted for by the analysts? They use the same design tools. They use the same fabs. They use the same processes to develop the chips. Yet ASSPs don't go into these analysts' totals for ASIC starts. So depending on which analysts you believe, ASIC numbers are either headed down severely or up severely. If it looks like a duck, talks like a duck, and walks like a duck, isn't it a duck?

Wally Rhines (Mentor Graphics): But the way this industry and all industries work is they move from customization to standardization. So the ASSPs have the ability to amortize the development cost over multiple users and multiple applications. And we go through waves of this. Certain standard configurations of product become usable in multiple applications, and we get a big productivity boost, and lots of people can build lots of volume in that period. Then we hit another tough point, where we have to do semi-custom products until we evolve the next ASSP. So, I think it really is a legitimate delineating definition.

Q: Does that turn your market, though, into a commodity market where you're fighting Altera on price the same as some of the other markets, or are you actually able to push upstream with that?

Wim Roelandts (Xilinx): We know that we can expand our market if we can produce the same capabilities at lower cost. So that's what is really driving us: the continuous innovation that allows us to lower the cost of our components. And it is working for us in the sense that today, close to 20 percent of our revenue is from the consumer market. That's an indication that even in consumer markets, people appreciate the flexibility.

What you need to realize is that Moore's Law is also a big deflator. Once you're in the market where the only thing you get is lower prices but a lot more capabilities, you become what you call commoditized. That is not the case with programmable logic. If you sell chips at $2,000 apiece, you're still far away from commoditization, and we can continue to lower cost for several more years before we're in danger of that.

But I fully support what has been said before. The only weapon you really have is innovation. If you don't innovate, you're going to die. It's as simple as that. That is true not just in technology, I think it's true in services too. There was a time when services were a safe business because they couldn't be moved out of the country. Well, that is no longer the case. So everybody will have to continue to innovate.

Q:But isn't your innovation really coming at the lower density level? I mean your jump into the consumer business has come at the lower density level and not the higher density level where you have gotten the most publicity.

Wim Roelandts (Xilinx): Absolutely. Five years ago, we were selling relatively small chips at exorbitant prices into low-volume industries. Today, with advances in technology, we really can be in the core of the design in the low-volume business, but what we have expanded into is the higher-volume business with these lower-density parts. Five years from now, we're going to be selling 10-million-gate devices into commodity businesses at $10 apiece, I can tell you that.

Q:How much power are they going to consume?

Wim Roelandts (Xilinx): One of the things that I'm a very strong believer in is that if you have an issue and you put some human brains on it, you'll find a solution. There are a lot of solutions to power. There are a lot of techniques that you can use to reduce power.

Jerry Fishman (Analog Devices): One of the interesting things you hear is this constant thing about innovation, but there really is a circular loop in companies. I think that's why you saw that even in the 2000 downturn, a lot of the best companies kept their R&D levels high. In previous cycles, when business really got crappy, everyone slashed and burned the R&D, because you don't have too many knobs you can pull to help you get the cost down. In the 2000 cycle, you saw that a lot of the best companies didn't do that. By and large most good companies kept the R&D spending high because they're all concerned about how to keep the innovation flowing. It turns out that if you don't do that, and a lot of your business starts to commoditize and your gross margins go down, then you can't afford the R&D.

If you have a business model that's based on keeping the gross margins up, innovating, and spending a lot of it on innovation, you can keep that circle going. But if the circle breaks at anytime, and you stop innovating and the gross margins start to go down because everything gets commoditized, you can't get out of that spiral because you just don't have the money.

Aart de Geus (Synopsys): The question is what happens when the music stops? At that point in time, I think it's very clear that those who have IP and a platform that works from which they can do rapid derivatives, they have spent their money wisely. All the others have blown their money on just doing chips. The derivatives are very low cost. That implies having a fundamental architecture, maybe some cores. Maybe it's a DSP core in your case or a Power PC core or what have you, something that differentiates. And you build around that as much as you possibly can because all the others are going to lose their shirts at that time.

Wim Roelandts (Xilinx): One thing that is changing, which I think some of you mentioned, is that the OEMs are designing less and less. Look at the majority of work we have to do in providing the development boards, application notes, and you name it, because they don't have the resources. You said they were squeezed. They are the people who have to let go of their engineers, but the work has not changed. They're trying to borrow more and more from the suppliers.

Alex Lidow (International Rectifier): I think that's good news.

Ned Barnholt (Agilent Technologies): I think there's also a more intense focus around customers. A lot of this is very application-specific, customer-specific, or industry-specific. There's a fragmentation going on. So you build platforms and you adapt them to different people, different industries, and different applications. It's still a platform, but it's got to be very customer specific.

I think the days of taking technology and throwing it over the fence are gone. Having very intimate relationships with your key customers, to truly understand where they're going, is absolutely essential to being an innovator. Nobody has all the tools nowadays, either the application knowledge or the technology, to do the whole thing. So you have to work very closely together with your customers and maybe even partners to build a competitive solution.

Q:How do you do that and justify costs? Doesn't that increase your cost of sales?

Ned Barnholt (Agilent Technologies): In some sense, focusing helps your cost of sales because you're not trying to be all things to all people. If you focus your R&D activities around some leadership companies, you can focus your R&D a lot easier than trying to serve everybody.

You have to be careful on the cost side and in who you pick as your partners and the customers you're going to work with. If you're going to pick a commodity company that's going to focus on driving costs down, then you're going to be under constant pressure for cost. If you work with customers that are really trying to innovate and be at the leading edge and do the next breakthroughs, then maybe together you can develop some of those next breakthroughs.

So, I think you have to be careful about your partners, but there's a lot of opportunities to work even more closely than we ever did before with customers.

James Truchard (National Instruments): I think that creates a key challenge because if you focus on that leader that's making the product that's going to be the commodity, you get in trouble. So you have to be cleverly looking ahead to say okay, where is the higher margin going to be? If you're just focusing on that leader that's going to be demanding a commodity product, you're probably going to be in trouble. You have to have a more sophisticated process for looking at who your customer is and who you should be listening to for the next-generation product.

Q: If you're in the cell-phone business, is Samsung your customer or is Sprint your customer? If you're in the digital-TV business, is Sony your customer or is DirecTV your customer? Sprint can kick you out of the cell phone quicker than Samsung can. And certainly if you look at DirecTV and satellite radio and on and on, in any business with a subscription model it's the service provider that's making the decisions—not the product manufacturer.

Ned Barnholt (Agilent Technologies): I think it's a combination of both. A lot of the service providers were caught by surprise by the success of the camera phones. Many of the cell-phone manufacturers drove those because they were looking for new ways to add features to their phones.

At the same time, the service providers are looking at how to lower cost, provide better Internet connectivity, and a lot of other things. I think you have to talk to your customers as well as your customers' customers to really understand where the markets are going.

Bob Swanson (Linear Technology): I think that's a good story about the cell phone. When Nokia and Ericsson were driving their costs down, Samsung, as an example, was trying to come up with a product that looked different, and they hit a home run.

Aart de Geus (Synopsys): They're really close to their customers, because one of the things they do is they bring in hordes of 14-year-olds. They give them the cell phones, and within two hours you know exactly which model is never going to make it.

Q:Who is in a better position going forward, the smaller companies or the bigger ones with lots of breadth in their product lines?

Jerry Fishman (Analog Devices): I think in certain markets you've got to have some sort of depth. If you want to be selling cell-phone chipsets, you'd better be doing $500 million a year or you can't afford the R&D. There are some of the very large verticals that you just need bulk. If you don't have bulk, you just can't afford the R&D.

But there's a zillion other applications. The analog guys for years have sold to 50,000 or 100,000 customers. That's where the real money is. I think it would be a mistake for any of us to just focus on the large ones, when there's 100 different market segments that are not very large. Actually, that is where you develop a lot of the technology that you eventually proliferate in some of the large verticals. But I think you need both.

Keith Jackson (ON Semiconductor): I think the nature of the consumer business is really in favor of folks with scale, because all the things we've been talking about tonight—the rapid change, the approaches you take with consumers with hit or miss models, and so on—it means you can't be a one-customer type of company because you're going to miss it. You may hit this one, but you're going to miss the next one. So you've got to have enough scale that you can look for those opportunities to provide the uniqueness, the innovation we talked about earlier, but also have the resources to provide that initial production—or you will get squeezed out pretty quickly.

Michael Marks (Flextronics): Back to my issue about commoditization. If you're going to play in the commodity market, you'd better be big. You'd better be really, really good. If you're not really big and you're not really, really good, you'd better not play in a commodity market.

James Truchard (National Instruments): For example, we have 200,000 customers with 25,000 different companies that we work with. Well, that's one approach. And the other is a few big customers. They're both good models. They're just different.

Q:Is North America stuck with those smaller markets? Is that our future in North America?

Michael Marks (Flextronics): From what perspective? From a manufacturing perspective?

Q:Right. And even from a brand-name perspective, regardless of where it's manufactured.

Michael Marks (Flextronics): Oh no. You've got guys like Hewlett-Packard making printers for 70 percent of the world market.

Ned Barnholt (Agilent Technologies): I think there's huge architectural and technological advantages we still have in North America. But clearly from a manufacturing point of view a lot of the low-cost kind of things are going overseas.

Alex Lidow (International Rectifier): I think there's the leading edge in certain key applications where as long as you're free to foot investing a lot in R&D, and you're constantly at that leading edge, you can make money with a smaller scale.

Q:We've talked about power management, but I know some of you have quite strong feelings about power waste in the larger sense and how big of a challenge that is worldwide. China is in a crisis situation because of power shortages.

Keith Jackson (ON Semiconductor): The greatest thing again about consumer applications is that in electronics, everything needs power. They may not all need 100,000-gate logic chips, but everything needs power. So the opportunity there to differentiate yourself by having the most efficient solution, whether that's size or energy absorption or whatever it is, there's just plenty of opportunities every time there's a new consumer opportunity.

James Truchard (National Instruments): Exactly. Things like hybrid cars, the high-power semiconductors.

Alex Lidow (International Rectifier): Something that Bob probably would really hate about this business is that the consumer doesn't want to pay for any of these savings. So you have to give it to them for free. There's a realization that once you get there, once you really understand that that's what you've got to do, then you can have some fun with it. You can make a lot of money delivering something for free. I'm really serious.

About eight or nine years ago, we focused on the washing-machine business. It cost about $200 to build a motor drive for a washing machine, and it saves half the energy. Well great. To the average consumer, that actually is about a two-year payback, but so what? Nobody wants to buy it unless there's a flashy front cover. Then they'll buy some.

But we have an opportunity to save about $95 billion a year on a worldwide basis on just three appliances—washing machines, refrigerators, and air conditioners—and that's the energy savings. You know what? We're not going to capture any of that. That's actually going to get plowed into society. What we'll capture, hopefully, is a lot of money for our analog ICs because people like the features. You like a washing machine that's silent, washes clothes better, and does it in half the time.

Q:Do you do any lobbying with the power companies to do rebates?

Alex Lidow (International Rectifier): I don't believe in that stuff. The bottom line is if you get the government involved in it, if you give rebates, you create these really screwed-up marketplaces.

I'll give you one example. This is my hot button on this. You can get a subsidy from the state of California for buying a solar panel for your home. That sounds really good, right? Well, the energy from that solar panel will never equal the energy it took to make the solar panel. So our state is subsidizing, or is actually aggravating, the energy shortage.

Q:Aart and Wally, does the IP business become commoditized over time?

Wally Rhines (Mentor Graphics): Well, there is quite a bit of differentiation, all the way down to highly commoditized IP today, so it's no different from other components. There's the very differentiated IP. ARM has an enormous business in differentiated microprocessors. There are commodity things that we all sell or provide in bulk to our customers, which are viewed as low value and in-between. There are standard-space IPs that are sold principally for the value of the support you give and the fact that if you're the largest or near the largest, yours has fewer bugs than other people's.

Aart de Geus (Synopsys): I think at the low end all the small pieces of IP are commoditized by definition, and it's continually moving away from complexity. The challenge is that for the higher-end IP—we're talking now about USB and USB 2.0, and from 1 to 2 it's a huge step, and from PCI to PC Express is a huge step—it becomes very difficult to verify or almost impossible to verify. So quality becomes an absolutely essential ingredient in this.

Imagine all the car manufacturers saying, "Oh yeah, I'm just going to buy a brake system from somebody." That IP needs to be verified so incredibly well that you can only do it if you sell a lot of brake systems to a lot of folks. So by nature the commercial IP will concentrate. The other pieces of IP will be in the hands of the guys that do the platforms, and that's the higher-end IP—the DSP cores, the processor cores, the memory cores—those are going to be captive because people build a platform around them.

Wally Rhines (Mentor Graphics): And interestingly, the market is segmented out that way. So you have the highly differentiated IP, sold by companies that focused on highly differentiated components. And the service-oriented IP, where being the No. 1 provider of USB 2.0 has a big advantage, is largely being handled by the EDA industry, where you know there's going to be support around when you need it. And you're going to have the same part sold to hundreds of different users. So it's going to have all possible applications tried out and debugged.

Q:At DAC [the Design Automation Conference] this year, there were a ton of small IP vendors.

Aart de Geus (Synopsys): No, at DAC this year there were a very few compared to DAC five years ago. Most have died. And the story is the same. It sounds great: Buy a piece of IP, four months later you have it in your chip. Your complex chip that's this big, the IP is this big, the IP is somehow not working, and you're totally stuck. Quality is absolutely a killer.

And by the way, the smaller companies are no longer around to help you, because after all you gave them only $40,000 to buy that IP. How long do you think they're going to stick around? The economics will work out to volume.

The interesting part is really within the larger companies. Can people start building their own IP for reuse? The efficiency within the design companies is in question as well. Initially, everybody says "Oh, I have a lot of IP. I'm going to just reuse it." But no project can reuse it with another project because there's no discipline, there's no standard and there's no verification.

Q:But the pushing of those standards has been going on with at least three different standard organizations, and it seems like we're no closer to it than we were.

Aart de Geus (Synopsys): Because standard organizations are primarily a forum for verbal interaction. A standard is something that's used a lot. That's what a standard is. Not something that's talked about a lot.

Earlier somebody was saying, "Hey, all you have to do is watch Intel. They're going to tell you what the standard is, and then you bet on it." Well, I know a few companies that bet on InfiniBand to the tune of $40, $50, $60 million—100 percent down the drain.

Q:You talked earlier about success being when you can push out the finish line. You're trying to control your inventory and your cost. How do you adapt quickly enough in this market to innovate while you're still trying to control the bottom line at the sa me time?

Ned Barnholt (Agilent Technologies): I don't think they're inconsistent. I think you have to be driving costs in your operations, in your manufacturing, whatever, and I think there's a little bit of a misunderstanding around innovation. I think innovation really applies across the whole enterprise—not just R&D. It can be in manufacturing. It can be in marketing and sales.

I think you somehow have to keep the innovative culture alive across the company while you continue to drive cost. But I don't see those as inconsistent. There's a lot of innovative things you can do to drive cost. If you can apply your innovation to drive cost, then you can get the best of both worlds.

Keith Jackson (ON Semiconductor): The other thing to keep in mind is that innovation is not a linear curve we're spinning. Innovation is all about the focus on solving problems. And getting a solution before the problem rears its head is where the real creativity comes in.

James Truchard (National Instruments): We were fortunate in that we didn't turn down as much as everybody else. Everybody else was turning down, so we could add more in R&D and spend more without looking bad relative to everybody else. So we got a little leg up there. Then we followed through on the upturn by saying, "Look, we're going to continue to invest in R&D at a level above what the upturn would give us in profitability if we did invest." And it worked in our case. The investors stayed with us. We didn't have to change our strategy from a high investment.

I think executives can do more to convince Wall Street and work the street to see that they understand the longer-term perspective. Because if you can get your investment out to five years instead of three years, it certainly helps your return in the long term.

Q: One of the key threads through the evening has been innovation. It sounds like the key challenge that you all face is how to create a company that innovates constantly. Because everybody else is going to figure out that innovation is important too. So what are some of the key ways you do that?

Bob Swanson (Linear Technology): You have to attract the right people.

James Truchard (National Instruments): It starts with the people. Second, for us over the years we've had a very intensive training of leadership in the key principles and key strategies that we want to work with. So, I personally spend some 44 hours on each class. It's a set of classes that we have, showing them how they can innovate and how they can look at the vision we have, the platforms we have, the product lines we have and innovate around our strengths. Almost always you have to start with your strengths and your core capabilities to be successful. The better the leadership understands those core capabilities and core products and platforms that you have, the better you can be.

Ned Barnholt (Agilent Technologies): One of the challenges for the industry, though, is that over the last several years so many companies have been hit hard with the downturn that there's a tendency to become risk-averse. I think what we need to do is go back to the roots of innovation—being willing to take risks, think out of the box, think creatively, not accept incrementalism. And I think that's a challenge for all of us. We need to remind people that that's been the strength of the industry. That's what's made us successful.

One of the things I'm concerned about, and I see it not just in our company, but other companies too, is that people have become a little bit more conservative because of the last three years. I think there needs to be a renewal of innovation. The spirit of innovation that's been so successful—such a key to the success of the industry.

Jerry Fishman (Analog Devices): You can't let Wall Street run your company. You sit and listen to these 26-year-old kids and they have never met a payroll, and they're telling you what you should be doing in the downturn? I think you learn that at the end of the day, when your business is good they'll buy the stock and when it's not they'll sell the stock and it doesn't much matter. Usually it's the other way around actually. At the end of the day that's what's going to happen. It doesn't much matter unless you have to raise money. So the goal is to never have to go to Wall Street for money.

Michael Marks (Flextronics): At the risk of being politically incorrect, which I tend to be all the time, you look around and there's just a bunch of old white men who are sitting here. That was my politically incorrect statement, and it's going to get worse. If you want to drive innovation, you'd better have a bunch of young people, and they'd better not all be from the United States. There's no Indians sitting at this table. There's no Chinese sitting at this table. There's no Ukrainians sitting at this table. There's no Mexicans sitting at this table. If you look at my company, we have explosive innovation taking place outside of this country. You have to absolutely harness the new forces that are evolving in the world. The companies that don't harness them are not going to be here 10 years from today when you're hosting this dinner.

Q:Does that vary by market?

Michael Marks (Flextronics): No, I don't think it varies by market. You have these developing countries that bring new cultures to the world, new attitudes to the world. You have people who are hungry. You have people who want to work night and day and who think that little victories are really big victories. I should add that we have people in San Jose who will out-innovate 500 people in India and 500 in China, so I'm not suggesting that this means that this is where all the innovation comes from. What I'm saying is that the world's changing. You have all these developing markets and incredible energy and resources and different ways of looking at the world. Integrating that energy and those resources and those capabilities and that way of thinking, which is different, is absolutely critical to the future of Silicon Valley.

I've sat around with some of my associates here in Silicon Valley saying "Woe is me, the Valley jobs are going away. We need to have government assistance." Government assistance? The world is a big place with fantastic people, and all these different cultures. Integrating these people makes for powerful, exciting, interesting and innovative companies. Some companies around here want to push that stuff away or ignore it or go back to the past. Forget it. These companies won't be here in 10 years. They just won't.

Aart de Geus (Synopsys): I'd like to second that. And the question is what has changed? Because in manufacturing that's been the case for a while. What has changed the high-tech business is two or three things. One is the skill level has come up to snuff dramatically. Secondly, communication is so prevalent that you can find out anything, anytime, from anywhere for free. And third, the resource distribution is changing dramatically.

In other words, with smaller resources in these new places, it only enables these talents. The thing that is still growing is actually the management capability worldwide. Both locally and how to manage globally, and you can see big differences between India and China. It took 10 years in India to get there, and most of it is managed through large multinationals. Whereas China, it's much more entrepreneurial, but also many more people have come back from the West. So for the companies that are sitting here, the question is, how do you do these things long distance? Mostly you start by screwing up and then you learn.

Q:All existing big companies try to innovate, but so many of the great innovations come out of startups—venture-funded startups. But so few startups have been funded over the past few years. Isn't that going to be a tough blow for the industry? There won't as many companies for a Cisco or a Broadcom or an Intel to acquire as there were a few years ago.

Wally Rhines (Mentor Graphics): Well, I think you'll have as many startups funded as there are good ideas to fund. If you look in the EDA industry, while it has consolidated, the big companies have picked up overall share. But for any given new technology, there are lots of new companies competing against the big companies, and they do quite well as a percentage. There were more than 80 venture-funded startups in our industry in the past year. So I think people with good ideas will get money. There are periods where it's harder and periods where it's easier, but if they can turn it into a return and if they can move quickly, then it will continue.

James Truchard (National Instruments): I think if you plot a line over time, you'll discover that there was kind of a bubble for awhile where a lot of crazy ideas were funded. Now we're back to a more normal way of funding good ideas.

Ned Barnholt (Agilent Technologies): I don't think it's a given conclusion that large companies can't innovate. I look back at my years at HP, and HP grew 18 percent per year for 60 years. And it didn't do that by staying in the same business. It didn't do that by keeping innovation at bay. It really concentrated on innovation. One of the things we try to do is decentralize a lot of the decision-making. You try to encourage people to take risks. We spent 8 percent of our R&D on far-out thinking about breakthrough ideas. We fund the venture funds, so we support some of these entrepreneurial companies. We realize we don't have the lock on all the great ideas. So I think big companies can still participate and drive innovation through a number of different mechanisms, not just leave it to the domain of the startups.

Alex Lidow (International Rectifier): I suspect that in the semiconductor industry the fabless model has spawned a very high level of startups. I really haven't seen it slow down. A lot of them go out of business, some of them get funded, and a lot of them get acquired. But they're happening every day.

Q: Is Wall Street more skeptical, more discriminating than ever? Is it a tougher marketplace?

Michael Marks (Flextronics): I did this talk on Wall Street just last week, and I had a big crowd, and I had a slide about inventory, because everybody was in a complete panic about inventory. I showed a 13-year history of what's happened with inventory, which is basically a straight line down. The revenue per inventory dollar has been improving for 13 years. Now they're all panicked about a little upturn.

So I say, "Here's this 13-year slide. For those of you who buy in the morning and sell in the afternoon, I realize you're going to have a little difficulty with this concept, which is time frame." But Wall Street is worse than it has ever, ever been in all the years I've been running a public company. There is an absolutely absurd lack of any kind of meaningful analysis. Wall Street has become totally driven by what's happened in this quarter. There's no point in talking about strategy. There's no point in talking about anything other than "Are my numbers going to be better this quarter than I told you or worse?" That's the only thing that matters.

James Truchard (National Instruments): Well the good news is you don't have to have three investors on Wall Street investing in your company, and I think that's a key point. You don't have to have every investor on Wall Street invest in your company. If you can develop a loyal following to see your point of view, I think that's key.

Q:This is the first year I've been in one of these roundtable discussions where there are no distributors in the room. The problems that have popped up and caused the recent dip we're going through have to do with inequities in the supply chain. So given the fact that there are no distributors here, will we ever get a handle on the supply chain?

Keith Jackson (ON Semiconductor): I think it's going to be difficult, but as long as we have the sell-in-recognition model, I think it's impossible to solve that problem.

We basically measure our products when they're consumed by the guys who are going to put them in boards. And that gets you a lot closer to a method of controlling that element of your supply chain. So, I'm a pretty big advocate for a different methodology there. There's always going to be irrational exuberance. Really what's important to our customers at the end of the day is making sure they have what they need when they need it. And us building too much doesn't help them. In fact, it hurts us in the long run because it adds cost to the whole equation.

You really have to have the toughest coupling you can get in the long run, to get rid of all the inefficiencies. All of that supply chain weakness, if you want to call it that, just adds more cost to the system.

Michael Marks (Flextronics): On the other hand, there is some lack of predictability in the world that people try to predict. People ask me all the time how is Christmas going? I say ask me in January. If you want people to walk into stores and buy stuff in the first week of December, you've got to plan that in October and November, and really nobody knows. They're trying to judge. All the sales takes place the last week before Christmas.

James Truchard (National Instruments): The double and triple ordering of components is what really messes it up, right?

Ned Barnholt (Agilent Technologies): I think one of the things that changed is the lead times have really shrunk. The uncertainty is higher. I think it creates a lot of issues. On the other hand, I do think that people are being more responsible. If you look at where we are today, you don't see the huge inventory overhangs that we had several years ago. I think people are trying to be a lot more cautious because they are waiting until the last minute to buy. They're waiting until their customers order before they buy inventory. I don't think we see quite the same buildup that we saw five or six years ago.

Bob Swanson (Linear Technology): We talked a lot about this last year. We talked about the OEM—this phenomena three years ago where OEMs came out with this huge forecast and took no responsibility for it. When Hewlett-Packard was a vertically integrated company, if some marketing guy said "Build a billion of these," and nobody could sell them, the manufacturing guy strangled the marketing guy. But now, the marketing guy is your customer, and he builds your product here. But I'm glad to hear that you think that some of this responsibility is coming back. When there was no risk at all for over-forecasting, just under-forecasting, then what happened was very predictable.

Alex Lidow (International Rectifier): There are some mature supply chains. Automotive is a good example of a mature supply chain. Even though there's somewhat volatile demand, you've got a very tight supply chain, very large accountability.

Bob Swanson (Linear Technology): Well that's because you have to have your parts in a truck circling around their factory. You can't afford to do that for every customer.

Alex Lidow (International Rectifier): That's true, it's tighter, but it's a 48-hour delivery.

Bob Swanson (Linear Technology): I don't know about you, but some of my products take me 16 weeks to make. So, how am I going to reconcile when they want 48-hour delivery and somebody needs to tell me 16 weeks beforehand what to build?

They get away with it because the suppliers said "I'll take responsibility for any bad forecasts." But when Dell says they want to do it, and all the big guys say it, pretty soon the thing is totally out of control, and we saw that three years ago.

Alex Lidow (International Rectifier): We saw certain US car models in April, May, and June go from 10,000 or 15,000 a week to zero as they corrected inventory, but there was no buildup in the channel.

It's a mature supply chain. It has its ups and its downs. But there wasn't this triple-ordering, frantic thing. When the car didn't move off the dealer's lot, that's when they stop sending you the pull signal—as opposed to 16 weeks later when they call and say "Maybe you want to slow down a bit."

Bob Swanson (Linear Technology): But that's a supply chain that's not as volatile as some of these other consumer things. Another thing is it's an industry that's been able to get the suppliers to say they'll take all the risk in return for big volume.

Q: In wrapping up, first of all, I appreciate all of you for coming. I'd like to go around the table and ask you all one last question: What can we expect from your company in 2005 and beyond? Wim Roelandts (Xilinx): If I could predict the future I would be a rich man and not be working. Most of the gurus believe that 2005 is going to be a good year for semiconductors. Assuming that the world economy remains reasonably healthy, I think we'll see a good 2005.

From Xilinx itself, we continue to push technology forward very quickly. So in 2005 we'll be producing 65-nanometer products. It looks like 65-nanometer is not going to be as difficult as people predicted a few years ago. So we see a lot of progress being made there.

I think the trend that I described earlier is continuing. We'll have to supply more of the IP. We'll have to do more of the design for our customers because there are fewer and fewer people doing it.

Keith Jackson (ON Semiconductor): To use an old phrase, the only thing we can guarantee is it will be different. There will be change. Hopefully, most of that will be realized in more innovative products for more of these consumer markets that we spent a lot of today talking about. But, really I think the opportunities over the next couple of years are almost uncountable with the proliferation of electronics now in all of these new markets like China and India, giving us real opportunities to grow and change and make ourselves more efficient.

Jerry Fishman (Analog Devices): My sense is that the world is getting to be a tougher place. While I believe that 2005 will be a good year, I think you'll see good companies getting better and bad companies getting worse. That's just inevitable. There's going to be a lot more price pressure than we've seen in the past because there's too many competitors out there that don't add value, they just drag the prices down. I think it's going to continue to be a challenging environment. I think the opportunities will be there for the companies that have invested and don't use what happened in the last three months to blow their companies up and do really stupid things.

There's so much uncertainty out there no matter where you go, that it is causing people to be very, very cautious. I think 2005 will probably be a good year. And if you've done the right things, you'll do well. If you haven't, it's going to be a tough place to be.

James Truchard (National Instruments): Well, we invested at National Instruments heavily in R&D, so we hope to see a return on that investment in 2005 as these new products come to market and are successful. We also see the convergence of design and test. The two now work more closely than ever together because trends like system-on-a-chip and telematics, where you bring many different technologies into one system, make it very important that you have a high integration between design and test.

Michael Marks (Flextronics): We're going to keep taking market share at the rate we have been, so by the year 2010 there'll be no other companies in my industry except for us. As you might have guessed from my comments, we're going to do much more design and much more component supply, which is where we believe the margins in our business. We'll have a record year of revenue, record year of profits. We think the future is good.

Aart de Geus (Synopsys): I think that 2005 overall is going to be tighter than '04 for the entire industry. As Jerry was saying earlier, if you're well prepared, you'll do okay, but it's going to be a little bloodier just by virtue of pricing. We're going to be under attack with a lot of capacity coming online.

From a technology point of view, today, we've already seen the first tapeouts of 65-nanometer. We've seen about 80 tapeouts of 90-nanometer. We'll see a wave coming on in 2005. From a Synopsys point of view, we'll provide for the first time a very integrated area-power-timing test and yield solution, which is going to be an absolute necessity to the design of these chips.

Bob Swanson (Linear Technology): I think there's a good chance that this growth that we're seeing in the last five quarters has legs to it. We all agree that the chip business growing 40 percent to 50 percent a year is not sustainable, so I think we're going to see some corrections. But I think there's a good chance, for the next two years at least, that there's more legs to this economy. My sense is that not only in the US, Asia, and China, but even in Europe, there are opportunities that are going to have some legs.

As for Linear, we're going to continue to provide the hard solutions or solutions to the hard problems that our customers are facing. I'm looking at two customers who are actually pretty good in analog, but in general, we're finding that more and more of our customers are challenged with solving their analog problems, and it's going to be our job to try to make those problems go away for them.

Ned Barnholt (Agilent Technologies): I think 2005 is definitely going to be a slower year than 2004. One of the things we're seeing here is we've come off a couple of tough years, and the semiconductor industry and semiconductor test have been growing at 40 percent, 50 percent, or 60 percent. And that's far beyond the secular growth rates of those industries. So I think the growth rates are going to have to come down to the secular growth rates, which for most of these industries are closer to 10 percent and 15 percent—not 40 percent. So there is a period of adjustment. I think there's a little bit of capacity digestion going on. Perhaps a little bit of an inventory digestion situation going on in some areas, particularly in some of the wireless areas. But overall, I'm still optimistic that 2005 will be a year of relatively modest growth, good growth overall. But not at the kind of growth rates we've experienced in 2004.

As it's been said before, the trick to all of this is what new companies can offer. Even during the downturn, while we've cut back a huge amount in our overall employees, we've been investing a lot in R&D and innovation, so we've got a lot of new products coming out, new businesses we're entering next year. So overall we're pretty optimistic.

Again, I think the key comes back to innovation. There's going to be no free lunch. We're going to have to work really hard. It's going to be very competitive, but you've got to fall back on your intellectual property and your ability to innovate.

Alex Lidow (International Rectifier): I think power management is a good place to be because we're going to need more and more power to do all the things we want to do—whether it be make our cars or washing machines more efficient or allow the computers to work in the same size box while delivering more. I'm real glad about that because that's all I know.

Wally Rhines (Mentor Graphics): As with the rest, I believe that the inventory situation we saw in July and August is a temporary one. We'll blow through that, and the industry will grow in 2005, and the scheduled recession for 2005 will not occur. On the other hand, I think it will be a year of moderate growth, but continuing growth.

As far as my own company goes, the one thing I feel confident of is that we will grow faster than the overall industry, and I think the industry will have a moderate to modest growth rate. We've projected ours. It's something like a 6 percent or 7 percent growth rate, and that comes just because we like to say we do what others aren't doing. We pick markets where there's opportunity to do things that are unique and always where we intend to get No. 1 market share. And there are some of those that just have enormous cost impact on our customers; the design for manufacturing, tests, resolution enhancement, and things like that, which are enormous growing markets. So I think we can take advantage of that and grow faster than the overall market. But probably 2005 will not be a year of strong double-digit growth. We'll have to wait another year or so for that.

Q:You guys were a great group. On behalf of Reed Business Information and Reed Electronics Group, thank you so much.