EIGHT KEY FORCES:
Aart de Geus
Chairman and CEO, Synopsys, www.synopsys.com
Chairman and CEO, Texas Instruments Inc., www.ti.com
Chairman and co-founder, Wind River Systems Inc., www.windriver.com
CEO and president, Maxim Integrated Products Inc., www.maxim.com
CEO, International Rectifier, www.irf.com
Chairman and CEO, Cypress Semiconductor, www.cypress.com
Founder and chairman, Analog Devices Inc., www.analog.com
Robert H. Swanson Jr.
Chairman and CEO, Linear Technology Corp., www.linear-tech.com
Although the state of the electronics industry is less than stellar right now, key figures in the business maintain a positive outlook—for today and tomorrow—on both their companies and the entire industry. Although business is tough, stocks are down, budgets are tight, consumer confidence is thin, and market forecasts are sketchy at best, none of this seems to faze these leaders. The big picture, as they see it, is huge, especially for the long-term.
"Despite today's market conditions, we see a very bright future for electronics, in general, and signal processing technology, in particular," says Ray Stata of Analog Devices. "The world is leveraging electronics in everything we do."
Tom Engibous of Texas Instruments Inc. couldn't agree more. "By 2015 at least, I think the electronics marketplace will reach $2 trillion worldwide. That's a trillion dollars of growth waiting to happen in a relatively short time," he says.
This kind of optimism also pervades the EDA sector. "In the next 10 years," notes Aart de Geus of Synopsys, "the pace will not slow down."
Jack Gifford of Maxim Integrated Products be-lieves that the immediate availability of data has made everyone more productive and efficient, which has been good for the economy. Wireless technologies and the Internet will continue to be important, claims Gifford, but "until a model is found to effectively sell bandwidth, both Internet and wireless growth will be somewhat constrained," he says.
"For me," says Jerry Fiddler of Wind River Systems, "the core problem is usage. What do people want to do, what do they want to carry, and how do they want to use it? No one knows what we're going to be using, or how we're going to interface with it. The biggest change over the next few years is that we're going to be in a world of hundreds of billions of connected devices that allow us to access and capture in-formation. And it's not going to be just computers. It's going to be much more pervasive than that. It's going to be embedded, pervasive intelligence, with information and communications."
Engibous thinks that growth over the next decade will be driven by the increasing demand for many mobile communications form factors, suggesting a "fundamental shift in the marketplace," he says. "Signal processing will become more important than microprocessing. Rapidly, signal processing is becoming the engine for every phone call, wireless multimedia connection, personal mobile device, and photograph that people take." With that in mind, Engibous says that Texas Instruments is now ramping new 300-mm, 130-nm copper production lines to bolster its position in DSPs. The company's engineers expect to continue to introduce hundreds of new analog chips each year.
What Linear Technology's Robert H. Swanson Jr. calls the "new respect for the analog IC opportunity" will be "reinforced as we go forward" with the convergence of computing, communications, and the proliferation of portable products. From a nontechnical perspective, Swanson says that the industry must grapple with the fact that it's taking suppliers and customers more time and energy to negotiate terms and conditions as the first hurdle to doing business. Often this takes priority over technology, quality, and service.
Cypress Semiconductor has set a goal for itself that calls for posting its first $1 billion quarter in 2005. To accomplish that, T.J. Rodgers maintains that his company will need to grow beyond the capabilities of its core communications businesses. The firm seeks to close the revenue gap through investment-type transactions with independent corporations that Cypress has an option to acquire but is not obligated to manage on a daily basis. One of its "gap" initiatives is Cypress Microsystems (CMS), a company that it formed recently to attack the $7 billion, 8-bit microcontroller market. CMS produces programmable system-on-a-chip devices that replace custom microcontrollers in a range of applications. Cypress has a contractual right to buy CMS when it achieves certain revenue and profitability targets. Cypress also recently formed Silicon Magnetic Systems (SMS) to develop low-cost, low-power, nonvolatile magnetic memory and has the same option to acquire SMS as with CMS.
Stata sees a transformation under way from "producing" companies to "thinking" companies. "I expect this to continue, especially for small and medium-size companies," he says. "Among semiconductor firms, the change is evident in the pervasive use of IC foundries, for example." Stata says that the lifecycle of the produced product continues to shrink, as does the lifecycle of the technologies inside. He believes that sustaining the pace of evolution will require a focus on innovation. Creating an environment where innovators can flourish and produce will become even more important in the future.
Clearly from a technology point of view, industry leaders believe that there's still much to do. "Thermal limits are threatening to slow down the pace of progress in computing technology and even rewrite Moore's Law," notes Alex Lidow of International Rectifier. "Lowering operating voltages for our digital computer and communications chips creates fundamental changes in the delivery system used to supply this energy." In the past, Lidow says that the delivery of power to the digital chip was an afterthought. Now, he says the need for extremely high currents at very low voltages (130 A at 1 V going to 200 A at 0.6 V in the next five years) is driving concurrent development of superior power-management MOSFETs, packaging, control ICs, and architecture. "All of these disciplines must improve by an order of magnitude if we are to prevent power management from being the roadblock to higher performance," Lidow says.
Noting that history often repeats itself with the call for smaller devices and more devices, de Geus signals his own warnings. Now, he claims that the smaller devices (0.13 or 0.09 micron) are bringing a slew of physical phenomena that previously could be neglected. Labeled under the general category of "signal-integrity issues," they require a specific approach: up-front avoidance in the design flow. "Failing to leverage this approach," observes de Geus, "will result in an overwhelming number of chips that will never reach the market."
As practical design sizes shrink toward 90 nm, de Geus says that automatic avoidance of signal-integrity issues, assertion-driven verification, very automated design for test, and systematic design reuse must be mastered to have a chance of completing chips that work on time. According to de Geus, increased design complexities and massively growing capital costs require tighter integration of the semiconductor value chain (designers, fabs, integrated device manufacturers (IDMs), intellectual-property (IP) providers, and EDA). "EDA is unique in this equation in that we connect to all other parties in the flow," he says. Complexity is a major issue, de Geus adds, requiring a smarter approach.
Moving forward, Rodgers points out that design and manufacturing will become the limiting factor for the semiconductor industry. "As mask sets escalate beyond the million-dollar mark, startups and smaller companies will find the barriers to entry much more formidable than larger, more established companies," he notes. The cost, he says, will be compounded by reworks, which will increase due to the more exacting simulation requirements of 0.13-micron and below technologies. At the same time, Rodgers expects the cost of simulation to escalate. "With over 40 simulation software packages presently in use, half of our design costs are now nonsalary expenses," he says.
Fiddler's take is that software and hardware are getting significantly more complex. "You have to move up the integration chain," he says. New devices, he continues, "can do so much more that you have to produce much more software for them. We have to provide literally orders of magnitude more lines of code now, and that will continue."
Despite today's market conditions and sketchy projections, these executives at least plan to be as entrepreneurial as ever, if not more so. "Maxim will always be an entrepreneurial company," says Gifford. "To win, we think it's essential to be entrepreneurial." According to Gifford, the company needs to be lithe and quick to respond to new technologies and changing customer requirements. "Our entire attention is focused on being useful and valuable to our customer—the professional engineer."
To position itself for the future, Linear Technology has increased its circuit design engineering staff by 16%. It also closed its oldest wafer fabrication plant while completing a two-year project of bringing online a new, more efficient, and more advanced analog wafer fab facility. In addition, the company reduced the number of process engineers working on new process development activities and cut the number of independent sales reps selling its products. It has moved to a direct sales force for the U.S. market and increased its sales activities in China. Moreover, in the past year, Linear opened a new design center in Burlington, Vt., expanding its design facility in the Boston area by 50%. "Even though the market has contracted significantly and the near-term outlook is less than clear," Swanson says, "LTC didn't cut back on investments. We accelerated them."
"We are always on the lookout for dislocations and new opportunities," says T.J. Rodgers. "Tough times require you to be more selective, but they also provide advantages to companies that are well prepared." For Synopsys, de Geus says, the recent merger with Avant and the acquisitions of inSilicon and Co-Design Automation are all moves that support one objective—to provide the best, lowest-risk, complete IC design solution.
As for new and developing trends, Stata believes that signal processing will be foremost, especially in communications and computers. "These industries will migrate from electrons to photons to realize higher speeds and greater communications capacity that are better by orders of magnitude. Despite the current malaise of telecommunications visionaries, photonics represents the next frontier for innovation and business development for the semiconductor industry," he says.
Gifford thinks that wireless local-area networks, the gigabit Ethernet, and 3G wireless phones will be important. "What has changed is some of the hype, or unrealistic expectations, for the ramp up and near-term adoption of some of these technologies," he says. He also expects desktop PC sales to decline until consumers have a compelling reason to replace them. "Convenience, through portability and integrated functions, will continue to be important, both to businesses and individuals," he adds.
Swanson looks for more portable product opportunities. "Lower power, less voltage, higher density, a smaller-solution footprint, and increased speed are the trends gaining momentum," he says.
Although these eight leaders offer different viewpoints, they share the same common vision of unshakable confidence in the electronics industry today and tomorrow.