What's in the Cards for Si?: Industry leaders prognosticate at SEMI's ISS - Capital Equipment - Semiconductor Equipment and Materials International's 25th annual Industry Strategy Symposium

Electronic News, Jan 14, 2002 by Jeff Chappell

PACIFIC GROVE, CALIF. -- Personal helicopters won't replace cars in the next 25 years, nor will five-course gourmet meals come in a pill.

Nevertheless, the semiconductor industry will likely change considerably within the next quarter century. But just what guise or guises the industry will take in the years ahead seems to be anybody's guess here at Semiconductor Equipment and Materials International's 25th annual Industry Strategy Symposium (ISS).

Several longtime industry veterans and leaders of capital equipment OEMs discussed what the next 25 years have in store for the semiconductor industry. Jim Bagley, Lam Research Corp. chairman and CEO; Neil Bonke, former chairman of Electroglas Inc.; Scott Kulicke, Kulicke and Soffa Inc. chairman and CEO; Ken Levy, chairman of KLA-Tencor Corp.; and Brad Mattson, former CEO and current vice chairman of Mattson Technology Inc., took part in the panel discussion.

The panelists agreed that trying to accurately predict the near-term future of the industry is difficult and problematic at best. They cited prognostications made several decades ago by industry luminaries, such as assertions that no one would ever want or need a computer in their home, or that gallium arsenide would make silicon substrates obsolete.

"I don't know anyone, with the possible exception of Gordon Moore ... who came close to predicting what would happen [today]," Lam's Bagley said.

Having acknowledged the difficulty of doing so, the panelists nevertheless offered visions of the future that sometimes conflicted with one another, even as they agreed that the industry tends to be evolutionary rather than revolutionary. Some suggested the industry might be in for some rather severe upheavals in the decades to come as the industry matures.

"The likelihood that ... we have a semiconductor industry that resembles anything like it is today is zero," Bagley said of the industry a quarter century from now.

The next 25 years will indeed be very different from the last 25 years, according to Kulicke. He predicted that silicon devices might become so ubiquitous that they reach a market saturation point, which, if they do, will precipitate dramatic changes as the industry enters a shakedown period.

The problem is not the technological extension of Moore's Law in the future, but an extension of the feasible economics of Moore's Law, suggested Brad Mattson. "I think we're already seeing the degradation of Moore's Law," he said. As semiconductor devices more and more become commodities, only the larger companies will be able to compete on a financially competitive basis, he added.

Electroglas' Bonke suggested that if one extrapolated the recent emergence of companies focused on intellectual property (IP), in the future companies in each area of the industry, including equipment suppliers, will concentrate solely on IP. OEMs will, in effect, become "fabricationless" equipment companies, outsourcing the manufacture of their equipment. He noted that tester company LTX Corp. is doing this now, outsourcing all but critical component assembly for its Fusion testers to contract manufacturer Jabil Circuit Inc.

But the semiconductor industry may not exactly be a brave new world 25 years from now. "Things don't change nearly as much as we think they are going to change," he said.

KLA-Tencor's Levy suggested that if one follows the trends currently established in the industry, lithography tools will have feature sizes of 10 to 20 angstroms, and electronic systems will incorporate 1,000GHz RAM in time for ISS' 50th anniversary.

"The question is what will we do with them," Levy said of these incredibly small but powerful devices.

This was a sentiment echoed by keynote speakers and W. Michael Cox, VP and chief economist for the Federal Reserve Bank of Dallas.

"We're leaving the information age. Information is cheap," Cox said. He's dubbed the coming era the Imagination Age, as people's imaginations catch up with the technology that is currently available.

In the nearer term, keynote speaker Michel Mayer, general manager of IBM Corp.'s Microelectronics division, suggested that it won't be design shrinks that facilitate the ubiquitous semiconductor device of tomorrow, but the integration of new materials that will facilitate faster ICs that consume less power and have broader capabilities. Faster processing speed is no longer the chief focus of the industry, Mayer said.

"The new environment demands much more. Speed is not the only metric," Mayer said. "As we move out of traditional computing ... power consumption and battery life are increasingly important," he said. "I'm convinced, beyond the dot-com bubble, there is a tidal wave of pervasive applications," he added.

Meeting the demand for these applications and surmounting the technology obstacles means integrating new materials, such as silicon-on-insulator substrates, silicon germanium, strained silicon, and copper and low-k dielectrics, he said. To successfully accomplish this integration, materials suppliers, OEMs and chipmakers will have to continue to work closely with one another and work out the IP and other competitive issues raised by close cooperation.