By Ron Leckie, Infrastructure Advisors
Nov. 17, 2008 – Each year, SEMI hosts its International Trade Partners Conference (ITPC) in Hawaii, a premier event that brings together top executives from the semiconductor, semiconductor capital equipment and materials industries to discuss global topics of mutual concern. For the last 24 years, hot topics have been discussed and debated here, often resulting in lasting relationships developing from the networking that occurs. This year was no exception — the controversial topic was when, or if, the industry should transition to the next generation wafer size of 450mm. Three presentations framed the issue.
Bob Bruck, VP of Intel’s technology manufacturing group and GM of technology manufacturing engineering, is one of the champions of transitioning to 450mm sooner rather than later. He puts the R&D cost burden for the equipment suppliers at just $6-8 billion — approximately half of what was spent by that same supplier group to transition to 300mm wafers, though most industry experts estimate that it will cost the equipment and supplier base more this time around. His $6-$8B estimate would end up consuming ~30%-45% of the supplier industry’s net cash flow for the next four years, but if the number is indeed much higher the cash flow drain would be substantial, since revenue won’t start to provide any return until at least 2015. Industry/government consortia and IC maker collaborations will help meet the ROI targets that need to be satisfied in supporting the 450mm transition, Bruck postulated, but an aggressive schedule is yet to be supported by the supplier base.
J.S. Choi, CTO and EVP at Hynix, presented a very strong and impassioned case on behalf of many chipmakers that the 450mm introduction should be delayed to reduce the accompanying huge risks. These risks include delays to ramp 450mm to acceptable reliability and productivity levels; substantial silicon wafer cost burdens, which will offset any 450mm cost benefits; and that the 450mm drive would starve investment in near-term 300mm improvements. Choi agrees with the suppliers that at an $8B construction cost per 450mm fab, there will be a very limited number of customers who can afford to participate. He shared data showing that at the same 57nm node, it was actually 36% more expensive to produce 8Gb NAND devices on 300mm wafers vs. 200mm. There is much to be gained by further improving 300mm manufacturing technology per the 300mm Prime project, he maintained, before considering a 450mm transition.
Charles H. Fine, professor at MIT’s Sloan School of Management, posed a key question: “Can the industry continue to drive down costs AND grow the overall pie?” In comparing the chip industry to others such as automotive, aerospace and airlines, he suggested that consolidation seems inevitable and that late entrants can step in to overtake stronger, older rivals. He cautioned that when dominant firms think they have a lock on the supply chain, they may become complacent and the supply chain relationships damaged. In citing the lithography example, Fine pointed out that the 200mm to 300mm transition did little for productivity, but the subsequent implementation of step and scan technology delivered significant productivity benefits Cost reduction has, and always will, enable chip demand, so the focus must remain on how to best achieve it; and collective/collaborative management will be critical.
For my part, the scenario playing out now is exactly in line with the white paper I wrote on the R&D funding gap for SEMI three years ago.
The supplier industry growth rates have slowed, restricting the availability of R&D funding, while the demand for higher technology at lower cost increases relentlessly. The only way to solve this funding gap is to collectively make smarter decisions on what projects will deliver a return to all participants, or to have dominant players with their own agendas fund custom R&D projects at their suppliers. This debate is not over, but very interesting data and insight was shared.
Ron Leckie is president of Saratoga, CA-based Infrastructure Advisors, an independent research and consulting firm for the semiconductor, equipment, and related industries. Contact: www.infras-advisors.com.