Lam's Bagley Ponders Industry Cycles, Education
-- Semiconductor International, 7/1/2000
 |
| James Bagley, chairman and CEO of Lam Research Inc. (Source: Lam) |
SI: You became CEO back in 1997, shortly after our industry experienced one of its worst downturns. Now that everything's headed back up, is there anything you plan to change at Lam, to at least ameliorate the effects of these cyclic ups and downs?
Bagley: We've made many changes. For us, the most important thing to control during a strong period such as the current one is fixed costs. This also means head count. Although we're committed to our people, we unfortunately had to take pretty drastic actions during the last turndown, including terminating a substantial part of our workforce.
SI: Was this solely determined by the downturn?Bagley: No. Much of it had to do with the fact that we entered the downturn with poor productivity. Just as an example, our productivity in the September 1997 quarter — which is the quarter in which I took over the company — was about $240,000 of annualized sales per employee. The last quarter that we completed, in the March 2000 time frame, was about $410,000 of annualized sales per employee. By the end of this year we should be somewhere around $450,000. This means the amount of fixed costs, as a percent of revenue, has been reduced substantially and our profitability increased.
SI: Is this a function of the staff reduction, instituting new QA procedures, what?Bagley: The list of changes we implemented is long. Principally, we focused on lean manufacturing and turning into a lean enterprise. Everything we try to do in the company is directed to providing added value to our customers, while eliminating waste and lack of productivity through changes to our business processes.
SI: You must be one of the few companies not pressed for people. Do you support H1-B visa increases?Bagley: We do — primarily because we cannot get enough domestic technical capability. We don't graduate enough engineers in this country, and we don't do enough to interest high school kids in technology. Regardless of efforts to improve them, math and reading scores are appalling, particularly when compared with those of some of our leading international competitors. We shouldn't compare ourselves to ourselves. We don't educate our people well enough, and the result is that we have very few — as a percent of the population — in this country who pursue technical degrees. Most of our graduate schools have high percentages of foreign students; the majority of Ph.D.s graduating today are foreign-born.
SI: Still, much of the world looks to us for innovation.Bagley: It is a peculiar situation. From a business perspective, as a nation, we're looked upon as the technical leader. But we're lagging considerably behind in the percentage of our population that pursues technical degrees and in just overall general educational capability.
SI: Is there anything else you are planning to change at Lam?Bagley: We're focusing considerable attention on our technical roadmaps for products, to ensure we bring them out sooner, relative to customers' needs, than we have in the past. This is very challenging and is getting even more so with every passing month, as we try to continue this very aggressive technology acceleration.
SI: In three to five years, where do you expect to see Lam?Bagley: We'll be substantially larger than we are today. The segments of the market we address are growing better than 50%, and we want to be able to grow faster than that.
SI: Do you think the technology Roadmap is too conservative? It would seem that some of the timelines are telescoping.Bagley: That's true. My perception is that when we didn't have a Roadmap, people worked along, advancing semiconductor technology — and we did a very good job of it. Then we started doing the technology Roadmaps and saying, "This is what we'll need in three, five, or more years." It's almost as if at that point we set a bar, and everyone believed he had to execute the Roadmap faster, not to be even with everyone else, but ahead of them.
It's as if everyone thought, "I'm supposed to be at 150 nm in 2002, but if I'm there in 2001 I'll be ahead of everybody and able to introduce new leadership products that no one in the industry can make except me, and I'll be able to grow my market share dramatically and get higher pricing." This reasoning puts tremendous pressure on us, as equipment manufacturers, to execute much faster than the Roadmap.SI: Essentially, aren't your customers asking you to stay ahead of them?
Bagley: Yes, to a certain degree you have to lead the customer in technology. However, a technology roadmap for the industry is a very complex integration issue, where we're looking at segments of the total process. There's a wide variety of companies, all trying to develop capabilities for segments of a total semiconductor processing capability. We don't have the same challenge as a semiconductor company has; they have to do 150 process steps, while we do maybe 30. So while we have a more focused view of what we need to do, there's no question we're executing the technology turnover faster than it has ever been done in history — and the acceleration continues.
SI: Surely, this cannot go on indefinitely.Bagley: Certainly, everything has limits. I'm sure this acceleration will become self-limiting. If we forecast the future based on what we've been doing in the past and assume this level of acceleration will continue, then there would come a time when we'd be changing technology at such a rate customers wouldn't have a chance to build plants to accommodate it. In turn, their customers couldn't assimilate the technology into new products fast enough. The whole thing is self-limiting; otherwise, you'd have a technology that would be so far ahead nobody could use it.
SI: Now that the fabless model is coming into its own, will the fabs themselves moderate growth by acting as a flywheel, so to speak?Bagley: In all likelihood, that's just what will happen. An inertia develops as you have more and more installed capability at a technology node, because you must realize a return on it. What the customer would like to do is build today 130 nm fabs to build 200 nm devices in, and have that fab live through several generations of technology evolution over a multi-year period.
SI: Isn't this more or less what is happening now?Bagley: In a way. From a technical capability, they try to make the fab they are building to be more and more advanced, which means our ability to continue supplying the equipment becomes harder to maintain. Let's say someone wants to build a 100 nm factory, because he wants it to last five years. We can't get 100 nm anything; we cannot get 100 nm lithography, resist, nothing. All we can do is design the equipment to reflect the direction we think 100 nm will take. But nobody has solved the resist problem, for example, so it could happen that when it is solved, our etchers and CVD equipment might be incompatible with it.
SI: In view of all this, what do you see as your growth areas, and how are you planning for them?Bagley: Our areas of interest are wafer cleaning, plasma etching and CMP. We'll continue adding technology and products, at least for the near term, in those areas. Further out in time, we may decide to enter other businesses if we believe it would be useful for us to be in those areas, whatever they may be. Right now, what we're trying to do is solidify our position in those three areas. Once we decide we're very strong competitively, and that we can manage that part of our business very well, we'll expand into other product areas.
SI: And how will it be done?Bagley: Through a combination of internal development as well as technology and company acquisitions.
SI: Any acquisitions coming up?Bagley: We're always looking for the opportunity to find technology we'd like to acquire. We always look at companies that would add to our three basic market areas.
SI: Does it seem to you growth isn't as easy as it used to be?Bagley: It's not so much that it is more difficult, as it's the fact that we want to have a wider array of products so Lam can become more relevant to the customer's fab. We want them to look to us for a wider variety of capabilities. Right now, growth for us isn't a problem; we have all we can reasonably handle.
SI: What should your customers expect from Lam in two years?Bagley: They'll find our cycle times are considerably shorter than they've ever been. We'd like to be at the point where our cycle time — from a customer identifying a productivity problem in the fab to our providing them with a solution for it — is shorter than anyone else's in the industry.
SI: What should we do to ameliorate the industry cycles?Bagley: The industry's already done something that'll mitigate the severity of downturns. Some has been industry-related; some has been forced on parts of the industry by the world banking community. One of the things that's different now is that it's going to be more difficult for a region of the world to put capital together at essentially no cost, to overinvest it in order to gain share. We've seen this in Japan, for example, and it lasted for some eight years.
South Korea took this same approach, where they had government capital to build additional facilities. This subsidization resulted in tremendous overcapacity. It appears this will no longer be acceptable to the world community. Restrictions have been placed on South Korea by the IMF as a condition of their loans, and since then the government has instituted many changes. They've tried to limit the ability within a chibol to move money from one operation to another, and the government is no longer directing funds to the chibols as it did in the past. Hyundai has streamlined its company, and its semiconductor business ultimately may be segregated from the rest and become a stand-alone entity. Samsung has moved away from its automobile business, which was clearly an unnecessary market entry for it, because it already had too much automobile capacity in South Korea. I believe we're in an era when governments no longer support that kind of unrestrained capacity growth, in order to pursue a sovereign agenda. If Japan and South Korea had not overinvested, the downturns we experienced in the late 1980s, early 1990s, and then again in the late 1990s wouldn't have been nearly as severe. This doesn't mean we're not going to end in a situation where we overshoot demand. Growth for semiconductors is periodically perturbed either because customers cannot handle technology that fast, or a new application that was expected to consume many chips arrives late. Either of these factors could dampen demand for semiconductors on a temporary basis.SI: Aside from these cycles, what would you like to see changed in the industry?
Bagley: Ours is a high-risk industry. It has the most expensive factories in the world of any product; there' the boom-or-bust cycles we've been going through since 1966 and cutthroat competition. Reducing semiconductor prices is a very positive move except when, like the Japanese and South Koreans, someone begins pricing well below cost and doing forward-pricing in hopes of accelerating cost reduction faster than their competitors. This doesn't work. Beyond this, I don't believe the industry has any major problem needing righting.
SI: As an industry, what do you think we should prepare for?Bagley: The rapid disintegration of the industry — the separation of semiconductor companies from the manufacturing side — will continue. This already has occurred in the assembly operation, the actual packaging of devices, with the vast majority of packaging being handled by subcontractors. I believe the wafer fab will continue this move as dedicated foundries continue to grow. As these fabs continue expanding and gaining strength, they'll level the differential between themselves and everyone else's wafer fab performance. Avoiding these tremendous capital costs will become even more of a requirement, but you won't be competitive unless you're with one of the top fabs.
We could see China becoming a much larger participant in the semiconductor industry than it is today. I expect reason will prevail and the mainland Chinese and the Taiwanese will arrive at some kind of pragmatic accommodation that will permit Taiwan to do much more in the mainland. If that happens, it will begin a process that will make China a leading manufacturer of semiconductors in five to 10 years.SI: Essentially creating more competitors?
Bagley: It's interesting that the U.S. has created its own competitors, except for a period of time when we weren't as astute in managing our own fiscal activities. We've created competitors and ceded market segments to them because we could take their output and use it better in a next-generation set of products. We have continuously evolved to an increasingly higher level of capability, and there is no reason why this shouldn't continue. •
— Alexander E. Braun
