Philips plans 300-mm fab as part of new manufacturing strategy By J. Robert Lineback Semiconductor Business News (10/06/00, 08:31:37 AM EDT)
NOORDWIJK, the Netherlands -- Philips Semiconductors plans to announce a 300-mm wafer production fab in the first half of 2001 as part of a strategy to become a more aggressive IC manufacturer, said Stuart MacIntosh, chief operations officer for the Dutch chip supplier.
Philips is now discussing the multi-billion-dollar fab project with potential partners, and it is reviewing possible sites for the huge plant in locations around the world, MacIntosh told SBN in an interview during the company's press briefings here this week. The move to build a 300-mm (12-inch) production fab comes as Philips increases its wafer-processing capacity by 60% in 2000 over shipments last year.
"The decision [of location] will be based on how you treat 300-mm. Is it just another engineering exercise [to ramp into volume], or do you believe the technology needs to be closer to your home base?" said the Philips Semiconductors senior vice president.
If the 300-mm tool set and processes are fully ready for volume production, MacIntosh said Philips could decide to locate its biggest fab yet in North America or Asia, but if not, the 12-inch wafer facility would most likely be built closer to the company's headquarters in Eindhoven, he added. Philips is now working with 300-mm development partner STMicroelectronics in Crolles, France, to set up a 12-inch pilot line (see April 13 story), but MacIntosh ruled out locating the new production plant in that city.
Other key factors in building the 300-mm fab will be Philips' choice of partners (if any) and its tool suppliers, which are taking on more development work in each new generation of chip plants.
For Philips, the aggressive stance in 300-mm production is in contrast with the Dutch company's conservative approach to capital spending and investments in the 1990s. Parent company Royal Philips Electronics N.V. of Eindhoven was unable to keep pace with the rest of the chip industry in the early 1990s because it was nearly bankrupt and fighting for survival. After a massive overhaul of Philips Electronics, corporate managers continued to keep a tight fist on spending in the late 1990s--especially in semiconductors.
"We invested very cautiously and were proud to boost that we grew faster than the rest of the industry and invested less because of our outsourcing [policies]," explained MacIntosh, referring to Philips' move to produce more of its devices in third-party foundries and backend assembly plants five years ago. "That worked in the period when the market was not growing very fast--in '96 to '98. Worldwide semiconductor revenues were increasing just 5% on average, and we were double that at 9-10%.
"But the downside of this came when the market rebounded in '99, and we could not react as fast as we wanted," he added. Philips painfully fell out of the top 10 chip suppliers rankings last year, based on estimates by a number of market research firms.
Not only was the slip in market share painful to Philips but it was also a problem for its biggest chip customers. Philips' tight production capacity also complicated the company's recovery from a wafer fab fire in Albuquerque, N.M., which shut down a fully loaded 8-inch frontend last spring for about six weeks and disrupted cellular phone shipments by customer LM Ericcson of Sweden. Philips did not have additional BiCMOS capacity to make up for the loss production in New Mexico.
But now Philips is aggressively expanding its process capacity for radio-frequency (RF) BiCMOS as well as deep-submicron CMOS logic. Early next year, Philips plans to begin volume production of Qubic3 BiCMOS chips in a wafer fab acquired from IBM Corp. in Fishkill, N.Y. The fab will also be tooled for Philips' next-generation Qubic4 process, which uses deep-trench isolation and offers higher levels of logic as well as higher frequencies.
More capacity has been added at the MOS4 fab in Nijmegen, the Netherlands. Another new 8-inch frontend has ramped into volume production in Albuquerque. And, Philips' most advanced 8-inch (200-mm) wafer fab--a joint-venture, called Silicon on Systems Manufacturing Co. in Singapore--has begun running engineering wafers in preparation for volume production next year with partner Taiwan Semiconductor Manufacturing Co. Ltd. (see Sept. 27 story).
"We have decided to be more flexible because of our size and the scale of orders coming from customers. So we must invest more," MacIntosh explained. "If we have a fab ready a couple of quarters before it is needed, that's what we are prepared to do, versus not being able to satisfy customers," he promised.
Consequently, Philips Semiconductors has been given the green light by corporate directors to spend more and take higher levels of depreciation on capital investments, said Arthur van der Poel, president and chairman of the chip-making unit. On Thursday, van der Poel outline an aggressive strategy to grow Philips Semiconductors by 54% to an annual revenue run-rate of 10 billion Euros ($8.8 billion) by 2002, compared to sales levels in the second quarter of 2000 (see Oct. 5 story).
The Philips chip president also said investments have been increased more than 200% to 2.0 billion Euros ($1.8) billion in 2000, compared to 650 Euros ($572 million) in 1999. In the next few years, Philips plans to keep its semiconductor investments at 2.0 billion Euros, he said.
"The difference [in investment strategies] is that we would wait to see what the installed base was booking before deciding to build a new fab," explained MacIntosh, but he added Philips will no longer wait before launching new projects. "Being in this growth phase is quite different. We have been a conservative company, and to make this change has been hard, but we have to have faith in our ability to perform and win market share."
As part of its plans, Philips is not backing away from third-party foundries and contract chip assembly. In fact, MacIntosh aims to keep the percentage of outsourced wafers and assembly work at about 20% of the company's total capacity in both manufacturing segments. "If it goes down to 15% or up to 25%, I won't lose sleep, but I won't want it to be 5% or 50%," he said. "At 5% you are not big enough [to demand attention from foundries], and at 50% you lose control of your own destiny."
The chief operations officer added, "When you are dealing with big companies, you have to be able to give them what they want, when they want it. We are struggling to do that this year." |