Chip production gear:
After three bad years,
how big will turnaround be?
By J. Robert Lineback
Semiconductor Business News
(01/04/00, 11:18:52 AM EDT)
Nearly everyone agrees that 2000 will be a big year for semiconductor capital spending. The only questions are: how big and for how long?
For now, the outlook is a moving target. Most analysts and industry managers predict that capital spending will go up in a range of 20-to-27% over 1999. But some forecasters are now revising their numbers beyond even this bullish forecast.
"A 30% growth rate is not out of the question and it could end up being a little north of that," says Richard Hill, CEO of Novellus Systems Inc., San Jose-based deposition equipment supplier. "Based on projections from our customers, this [outlook] extends into 2001."
Other key executives agree. "There are enough drivers [for ICs] that it may be a much bigger year than what people are anticipating right now," comments Mihir Parikh, CEO of Asyst Technologies Inc., Fremont, Calif.-based supplier of automation and mini-environment systems for wafer fabs. "It is certainly going to be one heck of a year!" he adds.
"Assuming the world economies hold up, [2000] should be a pretty good year," declares James C. Morgan, CEO of Applied Materials Inc., world's largest semiconductor equipment supplier in Santa Clara, Calif.
"There is a shortage of capacity in virtually every product area," he explains. In fact, the outlook is so bright that some chip equipment managers are starting to worry that IC growth could be too strong in 2000. This, they say, would result in a return to product glut conditions within a couple of years. "We'd rather see a long growth period instead of one of these quick jumps or spikes," declares Werner Rust, director of marketing at San Jose-based Silicon Valley Group Inc.
This "problem" came up in December at the Semicon Japan trade show, where some analysts were talking about chip capital spending jumping 40% in 2000. "One debate going around Semicon Japan was whether or not we were bringing on the next downturn quicker," recalls Pete Convertito, manager of strategic marketing for SVG's Lithography Division in Wilton, Conn. "Hopefully, that's not the case," adds Rust.
The recovery in capital spending began in the second half of 1999, resulting in a stronger-than-expected 10.4% jump in global equipment revenues to $29.2 billion, according to estimates by VLSI Research Inc. The San Jose market researcher predicts that revenues for chip production systems will rise 26.6% to $37 billion in 2000, followed by a 26% increase in 2001 to $46.5 billion. Another
That's not the end of the good news. "We now expect three very fat years," declares analyst Risto Puhakka, vice president of operations at VLSI Research. Chip equipment sales in 2002 are expected to grow 30% to $60 billion, he predicts, before they flatten out in 2003.
"Capacity is very, very tight, and it will only get tougher to keep up with semiconductor demand," Puhakka notes. "We're already seeing the new fab announcements starting to roll in, and we haven't found any nasty surprises to spoil the party."
There are several reasons why the outlook for equipment sales looks so good. For the past 18 months, chip makers delayed new wafer fabs and the expansion of existing facilities while they waited for clear signs of stronger growth in IC markets. But now capacity is getting a bit strained, with fab capacity utilization running above 90% for the IC industry.
Chip companies are now accelerating their device shrinks to boost chip volumes from existing fabs instead of building new frontend production lines. As a result, many existing fab lines are reaching their limits in terms of handling new process technologies.
As a result, chip makers will soon need to increase their investments in new tools, materials, and processes for 0.13-micron feature sizes, capital equipment suppliers believe. "We have a situation where we are seeing a pretty rapid push to tighter geometries," notes Morgan of Applied Materials.
"I haven't seen so many material changes in one block of time," he adds.
Larger investments for copper processing and new low-k dielectrics will be needed for next-generation interconnects. And at least a handful of major chip houses are expected within the next couple of years to begin pilot production with larger 300-mm wafers.
This kind of action will generate a huge potential for capital equipment suppliers. "I'm not sure I would call this a recovery as much as it is delayed investments," maintains Gus Pinto, senior director of business development for KLA-Tencor Corp., San Jose metrology giant. "We're seeing the whole range of activities from expansion of existing facilities to delayed investments in empty fab shells," he notes. "We are also hearing from customers about future 'greenfield' fabs."
Getting a head start on this next phase of industry expansion are Taiwan's pure-play silicon foundry companies, which moved fast to boost their capital spending late last year. "In 1999, the foundries didn't wait for the upturn to fully materialize before they started spending more money," notes analyst Bill McClean, president of IC Insights Inc. Taiwan's foundry companies--giants such as Taiwan Semiconductor Manufacturing Co. and UMC Group--are now investing in new technology to "the tune of $60 million per week," estimates analyst Robert N. Castellano, president of The Information Network.
But Scottsdale, Ariz.-based analyst McClean isn't looking for a big spike in equipment sales this time around. "I don't think we'll see the 30-to-40% jumps in capital spending like past cycles, but we will see a couple of good years in the mid-twenty percent range," he predicts.
Even the DRAM producers are beginning to hike their plant investments. After the foundries, the strongest segment for investments will be DRAM vendors, says Pascal Didier, senior vice president of worldwide customer operations for Cymer Inc., San Diego excimer laser supplier. "The only issue [here] is Japan, where you are still seeing consolidation [in DRAMs]."
Asian capital equipment markets began heating up last August and will remain the strongest geographical region until shipments begin ramping up in the U.S. this summer, predicts Michael West, business and strategic technology director for SEZ Group's U.S. subsidiary in Phoenix, which supplies wet-chemical processing systems. "Taiwan will continue to be strong," he says, "but I think a surprise will be Singapore, which is building up quickly."
Another bullish factor now is the wave of tool orders for volume production after "technology buys" for process development and pilot lines dominated the picture in 1999.
"In the past year, orders were mostly for high-end equipment--0.18-micron processes and copper capabilities--but now we are capacity driven with the recovery becoming broad based," observes Jerry Cutini, senior vice president of marketing and business development at GaSonics International Corp., supplier of wafer cleaning and stripping tools. "If companies start spending money to fill them up, there are [enough] empty fab buildings to make for a very good year," he believes.
Backlogs already are looking real good. Like many equipment makers, Microbar Inc. entered into 2000 with its best ever backlog of orders. The private-held supplier of chemical delivery and recycling systems is seeing its best growth in photolithography, advanced metal processing (copper), and chemical mechanical planarization (CMP), says Robert MacKnight, president of the Sunnyvale, Calif., company.
The recovery will get an additional boost in the second half of 2000 by 300-mm tool orders, say some fab equipment suppliers. In fact, chip makers' 300-mm roadmaps will provide an early indicator of the industry's health in 2000, says Andreas Toennis, chief operating officer for rapid-thermal processing (RTP) systems for Steag Electronic Systems GmbH, which is based in San Jose. "I think many IC companies are trying to get ready now," he says, "and we expect to see more interest in 300-mm orders for R&D and process development lines."
Also expected to grow significantly in 2000 are R&D investments in low-k dielectric insulators for interconnects. "Copper has enjoyed the limelight in the past couple of years, but low-k dielectrics will be grabbing a lot more attention in the next year," predicts David K. Lam, chairman of Tru-Si Technologies Inc., supplier of wafer-thinning technology in Sunnyvale. "It's not entirely clear [yet] which low-k technology will win in dual damascene processes, but . . . I'm sure some solutions will emerge," says Lam, who founded Lam Research Corp. in 1980.
When the industry does come out with a low-k dielectric with sufficient structural strength for IC processing and reliability, "it will take off like a rocket ship," says Novellus' Hill. New insulators for interconnects with dielectric constants of 2.9 are expected to be in high demand in 2000, he says, but "there are no leading candidates yet for the sub-2.5 to 2.0 range."
Even though chip production equipment orders are picking up strongly now, there are still those naysayers who worry about how such bogeymen as process shrinks would affect their orders rates.
"We have yet to see what the effects of process shrinks will be on future [equipment] demand," cautions Michael Wright, senior vice president of corporate marketing at Entegris Inc., a Chaska, Minn., supplier of fluid transport systems and wafer carriers. "With shrinks, a fab can increase volumes by 40-50%," he says, "and some IC manufacturers believe they will only have to change as little as 20% of their equipment to achieve these shrinks."
|
