Robotics may need jumpstart from 300-mm wafer
By Bill McIlvaine
What will it take to get the moribund market moving again for robotics wafer-handling equipment? Experts aren't sure. They're split now between it needing a launch of the stalled, next-generation 300-mm wafer or simply an uptick driven by the current generation of 200-mm wafer systems. Without a doubt, the robotics business for the chip industry was hit hard last year by the no-show 300-mm wafer.
"Until that gets resolved, there's going to be a lot of problems" for robotics vendors, predicts Robert Castellano, president of The Information Network. "There's a lot of wait and see on the part of the IC makers and that's really affecting factory automation," says the head of the New Tripoli, Pa., market research firm.
Other observers, however, don't believe the robotics industry will need the 300-mm wafer to get going again. "What we need is unit demand for equipment," bluntly says Clark Fuhs, who heads up semiconductor manufacturing analysis for Dataquest. He expects the next demand surge will be 200-mm driven. "We're waiting for capacity buys to drive the next major unit uptrend," he says. "It will happen in a very limited way this year, but the bulk will be next year."
Meanwhile, robotics system makers will have their hands full with another problem that has dogged them for years. Observers believe that chip gear makers will make heavier demands on their robotics suppliers for increased repeatability and reliability. "A couple of the big users came to us and said they've had enough," says Dick Dexter, vice president of robotics product management at Asyst Technologies Inc. in Fremont, Calif. "They said the same thing two years ago, but we believe they are serious this time."
According to one Asyst official, the company has been approached by a major fab equipment maker to replace the frontend robots from another vendor in an existing product line because the tool makers was dissatisfied with their reliability.
Brooks Automation Inc. in Clemsford, Mass., has been getting the same warning signs from Novellus Systems Inc. The San Jose-based deposition-systems vendor has had an established relationship with the tool company for its vacuum robots, but "if we didn't do it for them, somebody else would," acknowledges Mitchell Drew, vice president of marketing.
Researcher Castellano says this kind of thing doesn't surprise him. With the industry now taking a "breather," he says, tool makers have more time to consider issues such as reliability, uptime, and cost of ownership - all major concerns in the move to the 300-mm wafer.
Dataquest's Fuhs, however, believes that this kind of situation is "more of a design call than a competitive call. I do expect that at least in a limited way the 300-mm frontend design concepts may be applied to the 200-mm tool in next two or three years."
Even though chip makers are starting to increase their spending for upgrading fabs, the current market for semiconductor robotics is still in a holding pattern. Forecasters say the real business for new capital equipment will not come until next year at the earliest. Going ahead with 300-mm wafer production lines, something on which the robotics makers have spent so much time and money, is still three or four years away.
As for 1999, the year will be flat at best for the semiconductor robotics industry, say forecasters. And that's coming after one of the worst years ever for this automation sector. Lumping in robotics for producing flat-panel displays, a sector that was hurt even more than was the fab automation business, global sales fell a whopping 34% in 1998.
The worldwide merchant market for robotics hardware and software for semiconductor fabs dropped 18% from $358 million in 1997 to $294 million last year this after shooting up 27% in 1997, according to The Information Network. That market includes automated transfer tools, vacuum and atmospheric robots, vacuum elevators and platforms, as well as software.
When the captive production of robotics gear is added, the total market last year hit about $1.25 billion, according to researcher Castellano. Applied Materials Inc. accounted for half of that - around $600 million worth. The Santa Clara, Calif., giant, unlike other tool makers, builds its own robots instead of buying them on the merchant market.
Tokyo Electron Ltd. (TEL) accounts for the second biggest concentration of captive robotics, at about $150 million, according to Castellano. That leaves about $200 million in other captive semiconductor robotics systems scattered throughout the industry, he figures.
Despite the falloff last year in robotics sales, most of the vendors are well capitalized, claims Mitchell Tyson, president of PRI Automation Inc. in Billerica, Mass. "When [this business] was good, many [firms] went out and raised capital. Many of the mid-sized and even the small ones have the financial wherewithal to last though this [downturn]," he maintains. However, "the thinly capitalized companies will have a tougher time."
Brooks Automation was the market leader last year, grabbing a 27.9% share, which was up from 21% in 1997, according to Castellano. Revenues hit $83 million last year, down a bit from the $95 million it reported in 1997.
A distant second was Rorze Automation Inc., which was followed closely by PRI Automation's Equipe division and Kensington Labs. "In the wafer transport market, PRI Automation faces heavy competition from Daifuku," notes Castellano. "That market dropped 50% in 1998 and will be flat in 1999."
Brooks Automation recently expanded its robotics business by acquiring Hanyon Technology, a Korean supplier of automation software and systems integration to semiconductor and flat-panel display manufacturers in South Korea and Taiwan.
The fall in the semiconductor robotics last year didn't stop at least one company from moving into this market. Adept Technology Inc. is making a big push into semiconductor robotics. The San Jose company is a veteran maker of industrial robots and robot controllers for other markets such as electronics and appliance manufacturing.
"A lot of people have asked us, 'why did you wait?'" comments Charlie Duncheon, director of marketing. "We think we may have timed it just right."
Adept also is busy making deals and acquisitions. It is now acquiring BYE/Oasis, an Austin, Tex., specialist in contamination control that it will use to create a new division to focus on the semiconductor industry. And in July, the company will unveil a new set of atmospheric robots at the Semicon West trade show in San Francisco.
Also beefing up its robotic product arsenal is Asyst, which acquired Hine Design last year and is in now buying Progressive Systems Technologies, Inc., an Austin-based automation company, to get into reticle handling.
The Fremont, Calif.-based maker of SMIF pods and mini-environments had a bad fourth quarter. Its revenues dropped in half, going from $165 million in the fourth 1997 quarter to $84 million.
But Asyst's Dexter is looking for better than flat sales for the rest of 1999. "A year ago, a lot of our customers were buying inventory and a lot of them now have leftover robots," he says. "They're using them up, but slowly."
PRI Automation also is optimistic about the rest of this year. "We're seeing increased quoting and planning activity," says Mitchell Tyson, president of the Billerica, Mass., company. "A good sign is when [customers] say they'll move up the project. Now we're seeing some schedule pull-ins [and] a slight increase in bookings." But it will take two quarters to turn these bookings into higher revenues and earnings, he adds, so PRI doesn't expect to feel the impact of the recovery until late this year.
supersite.net |
