Robert Akins, Cymer Chairman, CEO

Robert Akins (Source: Cymer)

SI: Although you purchased ACX in February, this has been a disappointing earnings year for Cymer, as has been the case for practically everyone else in the industry. How do things look at present?

Akins: It has been a disappointing year, especially when you consider that, at the beginning of last year, the industry expected that 2000 and 2001 would bring continued growth, a strong and prolonged industry resurgence. There is no doubt that the current slowdown caught the industry by surprise.

SI: Why do you think that was so?

Akins: Past downturns have all been supply-side-driven. This has been an industry where, as we know, chipmakers rush to buy next-generation tools and build very expensive factories, all at the same time. This results in market overcapacity that impacts price, things come crashing down and then comes a recovery.

SI: This doesn't seem the case now.

Akins: No. This particular downturn seems to be driven by the demand side more than the supply side. There's simply a drop in demand for some of the market's more dynamic portions, which until recently have been the main drivers — in addition to PCs, the cellphone and wireless communications sector, as well as the Internet infrastructure. There have been many reasons for the decrease in demand, such as the demise of the dot-coms and a number of macroeconomic global issues. But it's different in character from past downturns in the industry. Also, it happened relatively unexpectedly compared to others in the past. This event has been the primary determinator of Cymer's financial performance, and that of virtually every company in our sector.

SI: Of course, the argument could be made that, at least with PCs and wireless communications, as an industry we should have seen the downturn coming. What do you think can be done to ameliorate these ups and downs now that we seem to be involved in a double cycle of supply and demand?

Akins: I don't believe there is anything we can do about these cycles. You can talk about supply and demand, but the fact is that they are greed-driven. As in many other similar, expensive businesses, this kind of cyclicality is simply inherent to the business. We've all heard the theories in past downturns as to why the next upturn will provide slow and much more controlled growth, which would be great. And what happens? The growth comes suddenly and there is tremendous overbuying, resulting in overcapacity and another downturn. I don't see this going away because, remember, if you build a factory today it costs you on the order of $2B, and if you're the first out with that particular chip product, you can pay off that fab in two or three quarters. If you are the fourth or fifth out, you may never pay it off. Trying to moderate or get rid of this cyclicality is a nonsensical thing to attempt. A single company couldn't do it, and I don't think that even if we all got together, we could do it.

SI: Does it appear to you that some markets — PCs, for example — are becoming saturated?

Akins: As far as PCs go, I remember that in the early 1980s there were concerns about the saturation of the PC marketplace. Everyone who needed one would get one and that would be it. Obviously, that didn't happen, largely due to companies like Intel as well as the PC-box retailers who redefined the roadmap for PC speed and capacity. Intel certainly played a major part in this, together with Microsoft, requiring more and more processing power to get the job done. This created the ongoing demand for more powerful PCs. Trying to predict what will saturate next in the PC market is like trying to predict where a balloon will explode when you squeeze it on one side.

SI: What about the Internet and wireless communications?

Akins: With the advent of the Internet and wireless markets, I think there was a belief in the industry that this was a great thing because instead of having just one principal driver — PCs — which used 65 or 70% of all chips, now you'd have three sectors about equally as strong in their demand for ICs. This diversity of demand would bring the desired stability. Now we're learning that macroeconomic effects drive those three sectors up and down together. So now, instead of having only one factor contributing to the industry's cyclicality, you have three! I don't expect there to be any kind of a stabilization effect by the diversification of the marketplace.

SI: Are marketplaces becoming saturated faster than they used to? It appears that it took the mobile communications and Internet markets less time to reach that level than it did PCs.

Akins: PCs were around for a long time before people realized everything they could do and the industry became truly dynamic. There was a learning curve for that market, particularly on the part of PC owners, to learn about the technology and its applicability to what they do before demand manifested itself in a significant way. By comparison, these other markets are relatively new. I believe that the Internet sector is driven by the dynamics of the dot-com phenomenon. In the wireless sector, it's true that anyone who wants or needs a cellphone has one, and they'll wait until they can get something of clearly higher functionality or evident superior performance before considering an upgrade. This might be better Internet access, high data bandwidth transmission capability, etc. Everyone I talk to seems satisfied with his handset until, again, he can get something compelling in the way of added functionality. Functionality will come, partly due to the industry's ability to produce far higher innate performance ICs that pack much more processing power and speed into a smaller, lower-power-consuming package. It is all interconnected. Whether it'll be G3 or what have you, is yet to be determined.

SI: You mentioned the cost of building a fab. In view of the intensification of the fabless model, do you think fabs might act as a kind of flywheel in the cycle process? They might not be too inclined to build next-generation facilities until getting ROI on current capital equipment in the present facilities.

Akins: That was proposed a year-and-a-half ago; that the diversification of the demand for chips plus the foundry model would dampen cyclicality in the business. First, one must remember that the foundry business is only about 15% of the chip manufacturing business, so it isn't yet a dominant factor in the overall dynamic of the chip manufacturing cycle — it would have to be much larger before having an impact.

That being said, the theory makes some sense — if enough of the business were produced by the foundries and they were behaving themselves, that might take the edge off peak demand capacity investment by other IDMs. However, on the other side of the foundry model you have very aggressive foundries aware that, to be effective and grow their business, they can't employ last-generation tools. That's why you see someone like TSMC, for example, with a lithography roadmap that rivals Intel's. They're forced into having to purchase large quantities of the most advanced equipment, and they are competing against one another to get the overflow of business from the IDM or fabless design companies. It's the same greed factor of trying to get as much of the business as possible, and it'll take a number of years (if ever) to sort itself out. It'll be a long time before the fabs themselves exert a moderating effect on industry cycles.

SI: Now that you have acquired ACX, what are your short- and long-term plans? Have you managed to digest your acquisition, and are the corporate cultures in sync?

Akins: (Smiling) No, the digestive process is still proceeding. Most companies would tell you that, under similar circumstances, it takes at least a year to get to the point where one feels he is on the power side of the curve in integrating a company the size of ACX into the corporate fold. We're still working hard at it, and the fact that we worked so hard to become a more process-driven company makes them a real asset to us to expedite matters. We've been working with them, defining a common set of processes, communication and tracking protocols, which have become very powerful tools in assisting with the integration. It will take us until this time next year before we can say, "Yes, the integration is complete. We are one company now."

SI: ACX brings a new technology to Cymer. You are obviously planning to be players in nanomotion?

Akins: Certainly! As a developer of nanomotion and nanomotion control technology, ACX brings a technology that is very different from ours. We believe that nanofabrication and nanomotion will first be a material business in the semiconductor industry's future, since by definition it addresses anything that occurs at 100 nm or below. In a year or 18 months, when Moore's Law moves everyone to 100 nm for full production, that's when we'll enter the era when technology such as ACX's will play a material role. In a stepper, for example, if you are trying to print 80 nm lines but you have uncontrolled or unwanted motions of 40 or 50 nm, you're going to fail. The old techniques, such as increasing mass to add stability or using integral differential feedback, have been employed in things like steppers or inspection and mask-writing tools, and they have done the job. We believe, however, that around 100 nm they're going to run out of steam.

ACX will not only enable us to play in that sector, but also in mask writing, mask inspection, wafer inspection, pick-and-place tools — any tool where you aren't processing the entire wafer uniformly, but rather stepping and repeating the process, and the accuracy and throughput of what you are doing greatly depends on motion control.

SI: Tell us about your technology roadmap.

Akins: We don't see ACX as an immediate revenue contributor. It's a long-term planting of seeds for us, as are some other efforts in the company. We're increasing our complexity to grow beyond being a one-product, one-market company — to leverage our success in lithography to other portions of the litho cell or into semiconductor production in general. So our roadmap is dominated by our light source for lithography business, which we believe will be an exciting business.

Right now, we're all caught in this downturn funk, but that shouldn't keep us from planning ahead. Right now, I cannot think of a single company in our sector with anything as close to the opportunity that we have. Remember that lithography used mercury bulbs for more of its history, and then about late 1996 we began seeing the first real insertion of krypton fluoride light sources to do the most critical layers. We saw our first quantum jump in the total market during the krypton fluoride generation. Now krypton fluoride is the state-of-the-art technology, and we have significant market share in that area. During the next upturn, in addition to krypton fluoride, fabs will need argon fluoride light sources to power next-generation scanners to do the 0.13 µm shrinks and, of course, the entire 100 nm generation will arrive. We'll be able to participate there as well, and mercury bulbs will be just a very low-level business. We expect a significant jump in the total available market for excimer laser light sources for those two wavelengths. Meeting this demand may give us a compound growth rate of 25 to 32%.

SI: What problems will you be facing in lithography applications in two to three years?

Akins: There are two. The first is managing multiple wavelength development and deployment. We aren't just a krypton fluoride light source company for lithography, but must also offer argon fluoride, F₂ and EUV — we must deal with the development and deployment of all four wavelengths. For the first time in the next year, we'll deploy all four, making this a much more complex business.

The second problem is that all these tools' overall performance must increase rapidly. Historically, the pulse repetition rate of our lasers has been the single biggest differentiator and challenge. We've built 1 and 2 kHz light sources. Recently, we introduced a 2.5 kHz krypton fluoride light source, and with our new argon fluoride source we now have a 4 kHz device. As repetition rates increase, the technology's difficulty increases exponentially. Developing and then manufacturing these very complex products in quantity, and then servicing and supporting them in a full production environment, can be demanding.

SI: Sometimes the importance of the light source gets lost in the shuffle.

Akins: Very true! People don't often realize that the light source is the first thing that must be made available to enable the development of next-generation lithography. Without it you cannot develop the optical materials or the photoresist that will be used. This is why our roadmap must be accelerated.

SI: How do you view the rash of consolidations in the industry — positively or negatively?

Akins: The days of new semiconductor equipment companies developing from scratch in garages to become large corporations have been over for some time. Semiconductor manufacturing has become a business where the large get larger and the small get smaller. We aren't the only ones facing significant R&D challenges to keep customers successfully abreast of Moore's Law. That's very difficult for small companies to do. Additionally, the demand by chipmakers for a global presence is mandatory for anyone who wants to be a player in this industry, and if you're too small you cannot manage it. Even if you're large, it becomes redundant for everyone to establish his own independent sales/service/marketing infrastructure on a global basis. There is a natural tendency to consolidate and leverage the existing investments across more product lines.

Also, most of the major companies are moving toward being total solution providers. As they span more of the processes and provide more of the solutions to chipmakers, they are in a more powerful position for success. This has also been a major driver toward consolidation. So we'll see more of it and, overall, it is a healthy situation.

SI: Is there anything you are planning to change in Cymer?

Akins: For us, process is the No. 1 contributor to change. As we decide how to carry out our business better, faster, more profitably, and more efficiently, we then reengineer the appropriate processes and make whatever changes in organizational structure or staffing that are required to implement that process. It's a never-ending cycle. Over the next five years there will be many changes in the company, driven by these business processes. In the coming eight years, I expect improvement to continue, although its rate will begin to slow down. With the acquisition of ACX, the introduction of our new product line, we're beginning to diversify. We've also introduced e-monitoring services to tie together our light sources in the chipmaker's fab, as one of the first steps into an area where we'll be able to bring more value, allowing the end user to concentrate on core competencies. We are planning our growth in an evolutionary fashion.

SI: So what do you do differently from your competitors?

Akins: We're different because, as I said before, we're a process-driven company. It wasn't always that way, but over time it has become our single most important competitive advantage. Although our competitors talk about other wavelengths, we're the only ones actually working on all four and expect to deliver them within the next year, proving our capability to commercialize technology.

SI: Which industry trends should we pay more attention to?

Akins: (Laughing) You can never pay enough attention to Moore's Law — that's got to be No. 1. When you build a fab, your budget for the production tools is somewhere between 75 to 80% of your total investment. Focusing on the ROI for those expensive tools is critical for chipmakers. Productivity and overall equipment effectiveness are becoming paramount in the end user's mind. Tools like EUV in the future could cost as much as $40M apiece. For this to be viable, the tool has to be very productive and reliable. A third dynamic we're seeing is the whole area of advanced process control and how tools can communicate and cooperate with one another, as well as the overall manufacturing equipment control system in a fab, to solve problems as they arise, maximizing process control and yield.

SI: What would you like to change in the industry?

Akins: In a free market, in a high-technology sector, I believe business does what it should be doing at any given moment. There is no way you can second-guess the infinitely complex equation that controls this industry. It's still one of the fastest, if not the fastest growing businesses in the world. With its ups and downs it may seem imperfect, but I cannot think of any suggestion that I believe would make it any better.