2004 Forecast: CEO Roundup
-- Semiconductor International, 1/1/2004
Advanced Energy Industries Inc., Craig JeffriesAdvantest America , Nick Konidaris
Amkor Technology Inc. , Scott Jewler
Axcelis Technologies , Mary Puma
Cimetrix , Dave Faulkner
Electroglas Inc. , Keith Barnes
FEI Co. , Steven Berger
FSI International , Jeff Butterbaugh
Genus Inc. , Tom Seidel
KLA-Tencor , Ken Schroeder
LTX Corp. , Roger Blethen
MKS Instruments Inc. , John Bertucci
Mykrolis Corp. , C. William Zadel
Nanometrics , John Heaton
Negevtech Inc. , W. Paul Shirra
Pall Microelectronics , Jon Weiner
Rodel Inc. , Nicholas Gutwein
Rudolph Technologies Inc. , Paul McLaughlin
SEZ Group , Kurt Lackenbucher
Surface Technology Systems USA Inc. , Andy McQuarrie
Ultratech Inc. , Arthur W. Zafiropoulo
Royce Instruments Inc. , Adrian Wilson
Craig Jeffries, Executive Vice President and Chief
Marketing Officer, Advanced Energy Industries Inc., Fort Collins, Colo., www.advanced-energy.com
It's always about the customer.
Providing value to our customers entails continually striving to make the key, benefit experiences that customers have with our organization better than competing alternatives. While current semiconductor industry issues (e.g., the ramp, 300 mm and narrower line widths, shifts to lower-cost regions, etc.) require attention, they can also distract us from vigilantly focusing on our customers.
Organizations often focus on providing "distinguishing features." Yet, continuously providing superior value to customers remains elusive. Why? Because "customer value" is often misunderstood.
Customer value is derived from the benefit experiences customers have with us, coupled with the price they pay, compared with competing alternatives. From the customer's perspective, benefit experiences provide superior end-result consequences. Customers often make trade-offs to find a superior "total package." We must learn about our customers' total lifecycle experiences with our organization, products and services as well as their experiences with competing alternatives. And it is often useful to begin with our "end" customer first. Once we understand how to improve our end customers' experiences, we can meaningfully affect key partners in the value chain.
As our industry matures, the companies that generate customer insights that provide differentiating value at critical points in the value chain are the companies that will be positioned to generate sustainable profits. And this will be a key trend to watch.
Nick Konidaris, President and CEO, Advantest America, Santa Clara, Calif., www.advantest.com
Driven by consumer electronics and communications tools, the long-awaited recovery, which began late last year, is picking up steam — many in our industry anticipate that it will last through 2005. It is becoming clear that our industry's cycles closely correlate with economic cycles. Automated test equipment (ATE), in particular, is one of the industry's bellwethers of upturns and downturns, with swings much more intense than the economy itself. Despite a global economy and easier access to information, it does not appear that future cycles will disappear or become less pronounced. Market psychology and human factors, together with ubiquitous technology and disruptive ideas, will foster new swings. Competition will continue to drive innovation in new applications and speed/bandwidth in existing ones.
ATE has experienced the recovery earlier than the rest of the equipment sector. The volume of devices to be tested has burgeoned, and with their increased complexity, the demand for longer test times is real. In the system-on-chip (SoC) testing arena, a disruptive business model is demanding one true, singular mainframe open architecture, with customized measurement modules representing the essence of ATE going forward. This business model is going to drive restructuring of the SoC ATE industry during the current upturn. I expect that in the following upturn (sometime in 2006 or 2007), the industry will see one dominant open architecture solution for ATE.
Scott Jewler, Senior Vice President, Assembly Business Unit, Amkor Technology Inc., Chandler, Ariz., www.amkor.com
While top-tier outsourced assembly and test suppliers (OSATs) continued to invest in technology through the downturn, IDMs had to cut back internal package development resources. IDMs are now forced to reevaluate their make-vs.-buy decisions. The continuing need for cost control favors strategically located megasites that can achieve optimal economy of scale in assembly and test.
Higher-performance silicon and systems are driving interconnect technology to increased complexity. The need for cost performance throughout microelectronics is driving package proliferation to achieve just the right performance level for each application. In advanced applications, interconnect's contribution to system performance is increasingly driving design complexity that demands collaborative and parallel device, package and system design efforts from the outset of any product concept.
Hot package types in 2004 will be QFNs, stacked-die packages, stack packages, flip-chip, optical packaging, and system-in-a-package (SiP). The leading edges of these technologies exist within OSATs rather than IDMs. Hot assembly process and materials technologies in 2004 include green and lead-free processing, and wafer-capable packaging for 90 nm, copper and low-k fabricated wafers. We will also see 35 µm inline wire bonding and low-cost high-density flip-chip substrates.
Materials, equipment and assembly suppliers are working to ramp capacity for 2004 growth. The need for multi-die packaging solutions is redefining relationships between design, foundry, OEM, ODM and OSAT players. The needed business models are being developed, but are not yet well established.
Mary Puma, CEO, Axcelis Technologies, Beverly, Mass., www.axcelis.com
Those of us in the semiconductor industry sometimes forget what a unique situation we are in — very few markets of any size can see revenues more than double over a two-year period, and then decline by 50% in the next few years. This is one unusual and unpredictable business, and if the last few years are any indication, this volatility is not going to change anytime soon.
The downturn and major industry changes (like consolidation and the shift of business to Asia-Pacific) have forced us all to evaluate our business models and find opportunistic ways of improving them in response to these market dynamics. It's a challenge, but it's also an opportunity to build a better business model.
To date, this challenge has led many firms to consolidate with or acquire other companies as a way of better leveraging R&D investments. It also has pressed semiconductor capital equipment firms to take a hard look at manufacturing models. For instance, many companies throughout our industry, Axcelis included, have pursued lean manufacturing initiatives in an effort to both reduce cycle time for customers and improve gross margins.
Moving forward, our industry will continue to find new ways to enhance profitability. Specifically, we see strategic sourcing and a more global supply chain as the next evolution of the lean manufacturing model. This includes outsourcing of some engineering initiatives — where supply partners take a stronger role in the design and build of subsystems or software.
This drive toward better business models serves us well not only in the downturn, but makes our companies stronger, and our industry more resilient and better prepared for another set of challenges — the ones we'll face in the coming upturn.
Dave Faulkner, Executive Vice President Sales and Marketing, Cimetrix, Salt Lake City, www.cimetrix.com
During 2003, we began to see a shift in 300 mm automation interfaces from "first installs" to "interface optimization." Equipment suppliers have installed their first interfaces, learned fab-specific scenarios, and are now discussing the optimization of their interfaces. Key indicators are SEMI's accelerating work on equipment data acquisition (EDA) standards and increased attendance at the 2003 APC/AEC conference, where we saw representatives from all major fabs.
Now that the first implementations of 300 mm are in place, fab managers are concentrating on how to get the most efficiency out of their company's investment. They are looking at ways to do this through better implementation of software and data-intensive applications that are particularly applicable with 300 mm.
This is important for the industry for several reasons. First, it gives our company an opportunity to invest in and build new products to support implementation of EDA standards and the ease with which they can be put in place.
Second, it is driving investment by many large process tool manufacturers — suppliers who are in the thick of the toughest 300 mm applications — to look again at their tool's software design in light of EDA standards. This will result in a wave of better tool utilization, better process feedback, better communications with host systems, and better materials handling.
Keith Barnes, Chairman and CEO, Electroglas Inc., San Jose, www.electroglas.com
The forecasts we have seen at the end of 2003 from SIA/SEMI demonstrate that there are a number of trends in the IT and consumer marketplaces that could drive a semiconductor equipment upturn for several years. This supports a positive outlook for 2004.
However, despite an anticipated upturn for the industry, it is my hope that the trend going forward for equipment and materials companies will be to assume practices that heed the lessons that we have been taught over the past few years. That is, we have to adopt more sober business models designed to maintain profitability or near profitability at the bottom of a cycle, then leverage resources for maximum profitability at the peak.
We must remain innovative, yet keep a constant focus on core technology, service and support — especially during down times — to maintain customer satisfaction, quality and serviceability. We must also adopt methods that will allow us to do a much better job of forecasting, including paying greater attention to non-traditional market indicators. Altogether, this can contribute to creating a more secure job environment for our employees and a steadier rate of return for our investors.
Steven Berger, COO, FEI Co., Hillsboro, Ore., www.feicompany.com
During the past two and a half decades, semiconductor researchers and manufacturers have been driven by the need for higher resolution, greater accuracy, and the ability to analyze and modify complex structures below the surface and in three dimensions. In the 1970s, the focus was on "microstructures"; today it is on "nanostructures."
We stand at the forefront of the nanotechnology age. Advances in nanotechnology will change the world we live in over the next 25 years — from breakthroughs in medicine and pharmaceuticals, to faster communications and rapid improvements in computers and personal electronics. The ability to view, measure, control and modify nanostructures in the fab environment is essential to this technological progression.
Fortunately, metrology technology has continued to evolve — and inline, automated NanoMetrology systems have become a critical category of equipment on the fab floor. Today's wafer metrology systems produce actionable information quickly, reducing cost, improving yields, and speeding time-to-market.
Yield-killing defects are more difficult to catch early in the process because they are now smaller and buried under the surface of several layers of an IC. Three-dimensional structural analysis has become necessary to provide quick answers needed to catch and correct complex process problems in the nano age.
Precise 3-D repair, editing and modification is a time- and cost-saving tool for making necessary changes to semiconductor masks and devices.
Managing structures on the nanoscale level is essential for enabling and managing naotechnology — and its expected impact and benefits to the world over the next 25 years.
Jeff Butterbaugh, Chief Technologist, FSI International, Chaska, Minn., www.fsi-intl.com
As much as we complain about the process challenges it presents, there is no stopping the technology trend known as Moore's Law. Lately, there even seems to be an acceleration of this trend. Indeed, Intel and other IC manufacturers have claimed they are now introducing new technology nodes every two years instead of every three years. This acceleration has moved the IC squarely into the nanotechnology arena with patterned features under 100 nm. Some may argue that IC manufacturing has already been in the nanotechnology arena, as some thin films have been used at thicknesses below 10 nm for several years.
In surface preparation, things may have finally reached the slightly absurd — or have they? The front-end surface preparation metrics for the new 2003 ITRS call for silicon and oxide losses of less than 0.5 Å for cleaning steps at the 65 nm technology node. This is less than half of the distance between a silicon atom and an oxygen atom in SiO2. Does this make sense? Is this measurable? Taken in the context of the entire IC manufacturing sequence, this requirement does make sense.
The increasing variety and complexity of transistors and circuits integrated onto a single chip are driving an increasing number of masking steps and thus an increasing number of cleaning steps. Currently, up to 20 implant mask steps are needed to form all of the transistors necessary for advanced logic devices. The number of implant masks will only increase with future technology nodes. Even at a level of 0.5 Å of material loss per cleaning step, several tens of angstroms can be lost throughout the whole sequence. This will represent on the order of 5-10% of the junction depth specified for the 65 nm node — and so does not seem so ridiculous in the end.
Tom Seidel, Executive Vice President and CTO, Genus Inc., Sunnyvale, Calif., www.genus.com
In discussions about the industry turning on or not, some are trying to analyze the emerging role of China. While it is still difficult to do business there, most are cautiously optimistic that China could have a significant impact. But what will China's role be in the industry?
I believe China will not accept a trailing-technology position, using only "shaken-out" processes. China fabs may start that way because it is practical, but it is not sustainable. China is starting with 150 and 200 mm wafers, but eventually the country will take on 300 mm wafers and leading process technology for its newest fabs, putting it at parity in design rules with most other regions.
The driver will be that any companies partnering with China, where they will invest in a new greenfield fab, will not want to invest in the current technology node minus one. This is investing in a process generation that will be obsolete right after the investment. Increasing capability in an existing fab would be less expensive. So, any technology partnerships that are going to catalyze growth in China will turn on with the current technology node.
Ken Schroeder, CEO, KLA-Tencor, San Jose, www.kla-tencor.com
For the past 40 years, the transition to successive generations of semiconductor devices has largely been driven by one key technology trend — shrinking design rules. Today, this is no longer the case. The insatiable appetite for better-performing devices by today's tech-centric consumers has led the industry to push optical lithography to once unimaginable extremes. The resulting technical challenges, combined with the looming threats of resistivity and electrical leakage, is driving the emergence of a complementary trend — the proliferation of advanced materials — to help fulfill the need for higher-speed, lower-power devices.
In 1970, five distinct materials were used in CMOS production, and it took 30 years to increase that number to 20 (at the 130 nm node). Today, the number of new materials in advanced CMOS production is growing at an alarming rate, and is expected to reach more than 35 new materials at the 65 nm node. Engineered substrates such as silicon-on-insulator (SOI) and strained silicon, new metal compounds and alloys such as cobalt tungsten phosphide capping layers, and ultrathin materials such as atomic layer deposition (ALD) barriers, are just a few examples of these exotic new materials.
Looking ahead, the industry will face significant challenges in integrating these new materials into production. Many will require radically new control methodologies to ensure their manufacturability. Even copper, which has finally reached near-yield entitlement levels at the 130 nm node, poses new and substantial yield and reliability challenges at the 65 nm node. While new process control solutions are needed, it is the way in which we look at process control that must fundamentally change. No longer will it be enough to simply take a measurement and look at each process step separately to evaluate its production-worthiness. It is the interaction of those process steps — how each affects the other upstream and downstream, and how each contributes to the overall manufacturability of the IC design — that must also be taken into account to ensure tighter production control. In the new era of semiconductor production, the axiom, "The whole is greater than the sum of the individual parts" is true. Only by looking at the whole process will chipmakers be able to reap all of the potential yield and performance advantages of these new materials.
Roger Blethen, CEO, LTX Corp., Westwood, Mass., www.ltx.com
The semiconductor sector showed good signs of recovery and growth in the second half of 2003, which is leading to bright prospects for 2004. Consumer markets, particularly video, broadband Internet, and wireless LAN will continue driving the industry's upturn. Digital signal processing has enabled new solutions throughout the IC market, and system-on-a-chip (SoC) implementations have continued to lower the cost of these chips, leading to a broad adoption of the technologies.
In the ATE industry, a fundamental technology shift to the scalable, single-platform (SSP) IC test system is being played out among several competing suppliers. Those ATE vendors that have adopted the SSP architecture should see market share gains and improving financial results. As the chip industry continues to recover throughout 2004, expect to see ATE vendors race to implement advanced RF test capabilities, gigabit-per-second serial data generators (for PCI Express), and innovative digital architectures.
John Bertucci, CEO, MKS Instruments Inc., Andover, Mass., www.mksinst.com
Our industry's supply chain has changed dramatically over the years, and critical components and subsystems companies have emerged to take ownership of process control technologies and solutions for improving productivity in the fab. As a leading member of this group, MKS is responsible for measuring, controlling, powering and monitoring critical process parameters.
Our business has changed significantly. We have expanded both our core technologies and process knowledge through acquisitions (10 since 2000) and new product development. We have become both suppliers and facilitators, responding to defined needs and proactively developing new capabilities that improve overall productivity for both OEMs and end users. We have come a long way in the past 25 years from our roots as an instruments supplier.
Many of the instruments, components and subsystems we supply to the industry are process-specific. Examples from recent MKS product introductions include a photoresist detection monitor for 300 mm cluster tools; MEMS-based vacuum transducers for wide-range pressure cycles in PVD and ion implantation; a gas analyzer to monitor the exhaust stream and provide real-time fault detection and automated APC functions on CVD, ALD and etch tools; and an ozone generator designed for ALD processes.
Today, our customers want meaningful information — not just data — for decision making and troubleshooting, and they want it in real time. MKS's web-enabled products provide actionable system information instantly from anywhere in the world over the Internet.
The next 25 years will require us to take on even more process and productivity responsibility and to continually increase our depth and breadth of process and process control knowledge.
C. William Zadel, CEO, Mykrolis Corp., Billerica, Mass., www.mykrolis.com
After several years of downturn, the signs are increasingly pointing to 2004 as the year of cyclical recovery for the semiconductor industry. Although the order trends may be similar to previous upturns, this one may be very different. After an extended period of unfavorable business conditions, many suppliers to the semiconductor industry are looking beyond industry growth and are focusing on how best to manage their businesses both in good times and in bad times. Improving operating performance by decreasing lead times and through careful cost containment are key priorities as companies seek to become more responsive to their customers while achieving and maximizing profitability in 2004. Companies will also capitalize on opportunities to augment their businesses, either through partnerships or acquisitions of strategic companies that can add profitability within a short time. These methods will help companies enter into rapidly growing subsegment markets and leverage their strengths in extensive worldwide sales and support networks.
Multinational expansion will continue to be a focus for companies striving to enhance market share worldwide, particularly in the Asia-Pacific region, and specifically in China, where opportunity is immense. Never before have major toolmakers and technology companies been exposed to a region with such potential. The sheer number of consumers — about 250 million — is breathtaking. These numbers, coupled with an extensive supply of hard-working engineering talent who have studied abroad at universities in North America and in Europe, could establish China as the ruling power in the semiconductor sector, with the lion's share of the $150 billion to $200 billion equipment and materials industry.
John Heaton, President and CEO, Nanometrics Inc., Milpitas, Calif., www.nanometrics.com
One trend that we will see taking on greater importance in 2004 and beyond is the emergence of integrated metrology based on the necessity for real-time process control. Process tolerances have become incredibly tight, and many tools have not been able to keep up. The only way to bring the process tools into compliance with these tighter tolerances is by incorporating real-time metrology. This is very difficult to achieve with standalone metrology — it requires the process tool to be interrupted and wafers to be physically transported to and from the metrology tool.
As a result, we will continue to see the increased application of integrated metrology using both feedforward and feedback process tool control in real time. Integrated metrology has already shown tighter control of key parameters on the manufacturing floor, enabling device and yield performance improvement, along with extended tool availability and improved utilization. Tighter process control also means lower material and process requirements. Integrated metrology also provides rapid fault detection, improved excursion control and loss prevention. Both high-performance and cost-effective results are needed for a broad range of ICs used in cell phones, cable modems, DSL, set-top boxes, graphics, wireless LAN, and voice-over-IP. These will be the drivers of next-generation testers and of volume purchases.
W. Paul Shirra, Vice President Marketing and Sales, Negevtech Inc., Santa Clara, Calif., www.negevtech.com
Clearly, fabs are coming back up to capacity and 90 nm production technology is at the leading edge. The next step will be pilot lines for 65 nm, for which we are already seeing planning stages, doing demos and development work.
The curious thing about any downturn is that Moore's Law marches on even when fabs are idle. No one wishes for a downturn, but in fact they are ideal times for developing new products for future production needs. During downturns, fab engineers tend to have more time to focus on new products, and how they help solve current and future problems.
Much of our view of the industry today stems from bringing up a new concept of wafer inspection starting back in 2000. During the recent downturn, we talked to many fab engineers and managers. They are almost blinded by the exorbitantly high budgeting costs for fabs and processing equipment. They are also very aware of emerging limits — both process capability and cost effectiveness — to some key processing areas. Our niche, wafer inspection, is one good example where many fab engineers believe that the conventional methods are ripe for fundamental change.
From what we learned during the downturn, we expect to see even greater emphasis this year, and with ongoing planning and development work for 65 nm production fabs, on further lowering fab costs for specific processes. It is no longer good enough just to make small incremental changes in cost of ownership (CoO); it is time for revolutionary changes that can change CoO by orders of magnitude.
Jon Weiner, Senior Vice President, Sales, Pall Microelectronics, East Hills, N.Y., www.pall.com/micro
For those of us in the microelectronics filtration business, the advent of new materials and processes has meant a shift in our focus — from old filtration models to a new model of filtration accompanied by purification.
To date, this has been most evident in gas filtration, where studies have shown that defect-killing particles aren't our biggest threat. Rather, molecular-level gaseous (H2O, O2, THC-total hydrocarbon) impurities pose a far greater risk to semiconductor yields, and thus must be removed to the sub-ppb level.
This trend is now surfacing in liquid filtration — and to a far greater degree. For example, solvents, developers and precursors for photoresists used in photolithography must be purified to the ppb level using reactive membranes that are selective for trace metal ions (dissolved iron, sodium, calcium and other ionic species contaminants) that will ultimately cause yield hits if left unchecked.
And this is just the beginning. Liquid purification standards will continue to evolve, becoming increasingly stringent, as shorter wavelengths and next-generation lithography processes demand newer, more pure materials. As we work to meet this challenge, we expect to see more collaboration between materials developers, toolmakers and device manufacturers, to understand the intricacies of new processes, to determine which impurities pose the greatest risks, and to find novel ways of eliminating contamination risks altogether.
Nicholas Gutwein, President and CEO, Rodel Inc., Phoenix, www.rodel.com
After a three-year slump, the second half of 2003 has given the semiconductor equipment and materials industry hope for a sustained recovery. Based on this momentum, 2004 looks to be a promising year for chipmakers and the entire semiconductor value chain.
No one "killer app" is driving this recovery. Instead, we are seeing a moderate, broad-based uptick in demand for electronics in consumer products, the start of the much anticipated refresh cycle for PCs — led by notebooks and wireless LANs — and a general recovery in the communications sector. Asia continues to be the fastest growing market, and it's not only based on strong foundries and Korean memory makers; Japan's restructured semiconductor industry is posting a phenomenal year of growth in 2003, exceeding 20% year-on-year.
We at Rodel expect this momentum to bring even stronger demand in 2004, and are working along with the rest of the industry to ensure that we can meet the accelerating demand for materials as our customers experience this moderate ramp. Like many in the industry, we believe that the silver lining in the post-dot.com bubble slump is that we will emerge stronger, smarter and ready to fully leverage this recovery.
Paul McLaughlin, Chairman and CEO, Rudolph Technologies Inc., Flanders, N.J., www.rudolphtech.com
Over the past 25 years, the same basic materials have been used for semiconductor manufacturing while dimensions have shrunk dramatically. However, we are now in a materials revolution that has become a major driver for metrology solutions companies. New technologies such as the sonar metrology in our MetaPULSE tools are solving new kinds of fabrication problems brought about by adoption of these new materials.
Interconnect is also undergoing change as manufacturers transition to low-k materials. The weaker mechanical properties of these materials present enormous challenges. Our metrology tools now provide unique solutions through non-destructive measurement of Young's modulus.
Transistor gates are also undergoing a materials revolution. Nitridation of ultrathin gates minimizes leakage currents and improves performance. Just a couple of years ago, XPS was the primary method for determining nitrogen content, but was far too slow for production. We developed unique combinations of laser ellipsometry and photomultiplier-DUV reflectometry to measure nitrogen concentration at production throughputs with process-control repeatability. Attention is now focused on high-k gate dielectrics and metal electrodes, and solutions are now being tested for both.
These are just a few of the metrology challenges brought on by the onrush of new materials. We expect new materials and processes to be introduced at an increasing rate. Metrology solutions companies can't follow. We must work ahead of manufacturers to provide the solutions they'll need to maximize production yields.
Kurt Lackenbucher, Executive Vice President and Chief Marketing Officer, SEZ Group, Villach, Austria, www.sez.com
The most important trend we see emerging is single-wafer processing overtaking batch processing. SEZ is focused on enabling full single-wafer capabilities while pioneering solutions in etch and clean, one of the last batch-technology strongholds. Long an industry goal, whole-fab, single-wafer processing will become a reality over the next one to two years as key technology drivers, including 90 and 65 nm device nodes, and 300 mm wafers, continue to emerge.
Geography is another key driver — technologies must be more globally available in the future, and companies looking to provide truly enabling technology need to have both a global and local infrastructure in place. Everyone is hoping the next industry cycle will be longer and more moderate; in reality, I believe it will be a rapid cycle, especially in Asia. Companies will again over-invest and need to digest the additional costs; this requires a financially solid company and structure.
I believe we’re looking at a strong 2004 and 2005 with great opportunities, but it will require a change in business model and approach — unlike the last cycle, in which business was equally split between the Western and Eastern hemispheres, 75-80% of business in the next cycle will be in Asia-Pacific and Japan. The key players have evaluated their 300 mm operations, so the big opportunities are almost preset; the next opportunity for equipment makers is to continue with R&D while managing innovation.
Andy McQuarrie, President, Surface Technology Systems USA Inc., Redwood City, Calif., www.stsystems.com
Markets peripheral to mainstream semiconductor manufacturing, particularly compound semiconductors and MEMS, which typically lag the major market by one to two quarters, are now showing signs of increased capacity utilization. One driver behind this is the market uptick for cell phones. While handset sales have been stagnant around 400 million units a year over the past few years, handset forecasts are back up with applications of picture phones as well as 2.5G and 3G applications. All this drives the compound semiconductor and camera device markets, which are two of our niches.
We are also seeing MEMS applied in more consumer electronics. For example, one dynamic consumer application is TI's DLP chip for projection televisions. In addition, manufacturers are increasingly adding MEMS accelerometers and gyros to toys and even laptops (i.e., the MEMS-based active protection on some IBM Thinkpads).
There is also a convergence of MEMS packaging techniques for mainstream packaging applications. This often involves the application of through-wafer etching (i.e., deep silicon etching) through thinned wafers to create bond pad connections that help reduce die size and overall package size. The net effect is an increase in manufactured die per wafer, this resulting in potentially lower manufacturing costs as well as improved form and functionality for the end user.
All these market and technology indicators bode well for STS as they rely on fabrication technologies that require the specialty plasma etch and deposition capability STS offers. Our business is definitely improving.
Arthur W. Zafiropoulo, President and CEO, Ultratech Inc., San Jose, www.ultratech.com
I believe that digital consumer products will be the single-most important driver behind the industry recovery in 2004. With that, 2004 should be an excellent year.
Specifically for the semiconductor equipment industry, it could be equally as profitable or better than 2000. My prediction is that we could see revenue growth for semiconductor equipment sales of 30-40% in 2004. Geographically, the United States should grow, but not as much as Asia-Pacific and Japan, which will both show strong growth. Europe will be flat to down, with a few exceptions.
In other areas, technology will play an important role as the industry moves to 65 nm technology. As devices shrink to 65 nm structures, there will be significant material changes required as well. In the past, there have been either device shrinks or material changes — but not both. Combined, this creates the most difficult transition in the past several years. On the horizon, some difficult challenges will need to be overcome before 65 nm can obtain the same yield percentages as 130 and 90 nm.
Adrian Wilson, Director Sales and Marketing, Royce Instruments Inc., Napa, Calif., www.royceinstruments.com
"Cautious optimism" seems to be the semiconductor industry's motto these days. At Royce Instruments, we base our optimism on three factors. First, from monitoring our environment, we have noted an increase in spending for capital test and assembly equipment overall, particularly over the past two quarters. Second, and perhaps our best indicator, we have seen significant and sustained growth in the level of sales activity.
Our third factor relates to what we see on our horizon. From the demand side, the horizon suggests improved economic conditions. Our industry now appears to be very closely in phase with the economy as a whole, which was not the case a handful of years ago. From the supply side, the horizon suggests further reductions of component inventory levels and an increase in fab utilization rates. This should give our largest customer segment — packaging subcontractors — the leverage they need to increase component pricing. This will, in turn, enhance their buying power.
Combining all three of these factors with our more aggressive stance from a new product development and distribution perspective, we feel comfortable with adopting the "cautious optimism" motto.
