Note the boldfaced section:
*******************************
Only CMP market was immune to plunge in '98 chip gear sales
By Bill McIlvaine
Last year's carnage in the semiconductor capital equipment industry hit nearly every one of its market segments, according to data just compiled by VLSI Research Inc.
The only area of wafer processing that didn't decline was chemical mechanical planarization," reports Risto Puhakka, vice president of operations. CMP revenues grew 5% in 1998 to $563 million, according to the market researcher. (see story in the May publication). Overall, VLSI Research figures that the global capital equipment market sank 25.8% last year, falling from $35.7 billion in 1997 to $26.5 billion. Breaking that down into major categories, sales of wafer processing equipment fell 30% from $19.2 billion to $13.5 billion, backend and packaging products dipped 30% from $3 billion to $2.1 billion, and test gear and software dropped 23% from $8.7 billion to $6.7 billion.
By sector, dicing equipment suffered the least, declining only 8% over its 1997 results, but every other segment was down more than 20%.
The results were far more grimmer for some market segments. Worst cases were diffusion gear falling a whopping 50%, ion implantation systems dropping 43%, computer integrated manufacturing (CIM) software dipping 45%, and wire bonders down 42%.
The brutal sales results hit some companies especially hard. Silicon Valley Group, which was the No. 4 stepper maker in 1997, didn't show up among the leaders last year. Tokyo Electron Ltd., which accounted for half the diffusion furnace market in 1997, slipped to 35% of a far smaller market last year. ESEC's S.A.'s die bonder business dropped from 30% to 25% of a market that was off 36%, while Toshiba Ltd. dropped out of the top five suppliers.
There were a few vendors that climbed the lists of leading suppliers last year. SCP Global Technologies Inc. (formerly Santa Clara Plastics Inc.) moved up among the leaders in wet processing gear and wafer cleaning equipment. And Promis Systems, now part of PRI Automation Inc., appeared for the first time in CIM software.
One surprise in last year's data to VLSI Research's Puhakka was Teradyne Inc.'s move to the No. 1 supplier of automatic test equipment. The Boston company's share of this market grew from 25% in 1997 to 30%. Also sporting rising market shares were LTX Corp. and Credence Systems Corp., he notes.
Varian's Aurelio
While the ATE market declined 21% last year, it was one of the better performing equipment markets, Puhakka points out. "The memory ATE market declined a lot - no question," he notes, "but there was an increase in logic and mixed-signal test." Teradyne, Credence, and LTX all gained market share at the expense of Advantest Corp., which saw its core markets in South Korean and Japan decline sharply, notes Puhakka. "I'm pretty sure that Advantest is going to bounce back once the spending in Japan and Korea comes back."
But it is going to take a lot more time for the entire semiconductor capital equipment industry to bounce back. It is on target to make a strong showing for the rest of this year, maintains Puhakka. Monthly shipment data, he says, shows that "the recovery is really sharp - a 'V-shaped' recovery." He expects this trend to continue, with December's shipping rates 40% higher than last December.
But continuing even that fast pace would result in only a flat year compared with 1998, Puhakka notes. "So that basically means [we'll have] a slow recovery," he predicts. "As a matter of fact," he adds, "we don't expect the industry to be the same size as [it was] in 1997 until 2001."
That glum assessment was echoed by other analysts.
The chip industry over-invested in 1996 and 1997 and is paying the price for it now, says analyst Mark FitzGerald, who covers the semiconductor equipment industry for Merrill Lynch & Co. from San Francisco. "We never should have spent $40 billion in the last cycle. It was an aberration," he says.
He predicts that industry sales will hit an estimated $27 billion this year, up from about $25 billion in 1998.
Spending won't reach the $40 billion range again until the 2002-to-2003 timeframe, FitzGerald predicts, or when 300-mm fabs start going operational.
By that time, however, there may be a lot fewer players in the semiconductor equipment business.
The depression that hit the industry last year already has permanently altered the landscape. The list of mergers is a long one: SpeedFam merged with IPEC, Steag with AG Associates, FEI with Micrion, and FSI with YieldUP. In Japan, Sugai merged with Sankyo to form S.E.S., which moved the combination into second place in the wet processing equipment market.
"In every succeeding downturn, fewer players emerge," observes Howard Silver, vice president of sales and marketing at M3 Technologies Inc. The Stamford, Conn., manufacturer turns out wafer chucks and handling systems for wafer processing systems.
"We're definitely moving in the direction of fewer vendors," says Merrill Lynch's FitzGerald. "It's been going on since the 1980s," he says. The capital equipment industry will continue its trend toward just two or three dominant players in each market, he predicts, with the remainder of the business split up among a bunch of smaller niche vendors.
That's not how one equipment company president sees it, however. Richard Aurelio, CEO of Varian Semiconductor Equipment Associates Inc. in Gloucester, Mass., doesn't believe the chip gear industry has entered a major period of consolidation. "What amazed me [this past year] was how little things changed in terms of the number of suppliers," he says. "Here we are in one of the worst downturns, and in truth there was very little consolidation. It speaks of the survivalist nature of smaller companies," he figures.
On the other hand, Aurelio believes that it will become impossible for the smaller vendors to serve chip makers without a global presence. He wonders how many will survive the next downturn.
But if these smaller players execute well, they can survive - and even thrive in niche markets, maintains VLSI's Puhakka. "I don't think it's an issue of size," he comments. "Once you reach any size where you can achieve global service and support, which is what [integrated device makers] ask for now, then it's an issue of how well your product works relative to that of [other suppliers], and how well you market it."
But other pressures are building up on the smaller suppliers. "What has changed [since] the downturn is the amount of critical mass that's needed to [run] a business," notes Aurelio. "Now it's virtually impossible to get into this equipment [market] as a startup. If you don't have critical mass," he says, "it will be virtually impossible to stay in the business."
One shadow that's spreading over more of the equipment business now comes from the leading supplier, Applied Materials Inc., which still seems to be growing market share. For the three months ended May 2, the Santa Clara, Calif., company reported $1.4 billion in new orders. That is estimated to be nearly 40% of the orders for the equipment industry.
For the first quarter of 1999, a period when industry orders hit a low point, Applied accounted for an impressive 50% of industry bookings, according to analysts tracking capital equipment. This does not include backend assembly gear or automatic test equipment, two markets in which the company does not compete.
Applied's improving performance showed up last year in deposition systems, the industry's biggest market. Even though the total business here sank 33%, the industry leader gained about 10 percentage points in market share, to 45%, according to VLSI Research.
Puhakka is effusive about the company's performance here. "Applied is just so good at developing those systems and their execution is close to perfection," he says.
But he points out that Applied has competition that it should worry about. Take CMP, where SpeedFam International and IPEC merged last year to fight off the Applied steamroller. "If SpeedFam-IPEC keeps it up and does what it's announcing," he says, there's no doubt this company will "give a good fight and maybe even reverse the trend. But," he adds, "it will need to be able to execute."
Another market where Applied should worry is the interconnection segment, Puhakka points out. Here, copper deposition is changing the process from physical vapor deposition (PVD) to electroplating and some new approaches. "PVD has been Applied's stronghold and now it's up for grabs," he maintains.
Market forces will prevent Applied from achieving the kind of dominance that Microsoft or Cisco has, according to Merrill Lynch's FitzGerald. That's because semiconductor companies are "better educated customers," he says, who know their technology and can differentiate among vendors. Chip makers "want the market to be competitive and don't want a sole source for technology," he says.
"Major customers [also don't] just look at your equipment and best technologies, but they also want to look at your balance sheet as well as your market," says Varian's Aurelio. "They are concerned about your staying power for the next 10 years."
Another problem that should disappear shortly is price cutting. The heightened competition during the downturn forced some discounting by suppliers. Some residual price cutting is still going on, says VLSI Research's Puhakka, but he doesn't expect to see such pricing pressures continuing for much longer.
"As companies introduce next-generation equipment, they're going to get normal prices," the market researcher believes. The "very strong demand for new capacity," he says, "will make pricing a critical issue. Prices are probably going to shoot up when an equipment shortage comes into play," he says.
But managers and industry observers alike believe that it will take a concerted move to 300-mm wafer processing to get the semiconductor capital equipment market back on its old growth curve. The higher productivity delivered by the new 300-mm tools also will be reflected by higher prices, but analyst FitzGerald still believes there will be a bit of a fight.
"Semiconductor companies recognize they're asking for equipment that's going to do a lot more," he says. "They just want to be sure that some of those advantages accrue to them. That is going to be battled over, tool by tool, with ASPs."
Intel Corp, which recently restarted its dormant 300-mm fab project in Hillsboro, Ore., holds the key here, FitzGerald says. "That's one of the things that's held [300-mm up] - Intel wants a cost-effective solution." Intel's goal is to get 12-inch tools that are only 30% more expensive than the 8-inch tool set, he says. "Whether [the vendors can] deliver it at these price points is questionable."
supersite.net |
