By Ron Iscoff, Editor
After a lean 1 1/2 years, with capacity going begging everywhere, the worst appears over for most IC packaging foundries. As usual during a recession, the weak get weaker-and typically drop off the map-and the strong survive to fight another day. As Dr. Subash Khadpe reports in this issue (page 52), consolidation has taken its usual toll. The dream sales and profits that were so evident at the end of 2000 for packaging foundries turned dismal by the first quarter of 2001. Last year the bloom was definitely off the rose. As we've seen in past down cycles, prices have dropped precipitously, and long-term customer contracts have suddenly vanished. Now, unless everyone's getting (and sending) the wrong signals, packaging and test providers are due for a healthy return to profitability. The head of one large packaging foundry recently told us that his company "lost" some 18 customers during the slowdown, due mostly to consolidation. "Now," he reports, "we're starting to see them return as customers, this time as subsidiaries of their new owners." Dr. Gerald "Skip" Fehr, executive vice president and CTO at OSE USA, San Jose, notes that the average selling price has not necessarily fallen during the recession. "However, that's only because the newer packages, i.e., flip chip, bring in more money. But when you look at the price of a common package a year ago and now, that package is cheaper today." This article's illustrations, by the way, show that despite many changes in the technology, the overall look of a packaging foundry hasn't changed very much over the past two decades. A Time for Expansion A few companies, namely giant Amkor Technology and ASE, built up their brick-and-mortar resources during the recession either by physical expansion or by acquiring smaller competitors. Amkor, which seemed to be riding hell-bent for leather on the expansion trail, began its entrance into Taiwan and added smaller competitors in Japan. ASE substantially expanded its Taiwan facilities and made ready to jump into mainland China, pending Taiwan's approval. For decades a packaging foundry's ability to offer same-site test was met with little excitement. Now, perhaps sparked by STATS' determination to buy as much test gear as the Singapore-based company could reasonably acquire, competitors are adopting a similar "we'll test it right here" mindset, too. Incidentally, STATS also bolstered its onshore test capabilities with a wholly owned subsidiary, FastRamp, which opened in Northern California in March.
Over the past 1 1/2 years, we've also seen a determined move to China by most of the top packaging foundries. This will continue. For the near term, it will mean moving old packages to China and keeping the newest ones at headquarters sites. We're also seeing a gathering of prov-iders who are looking at small, specialized packaging niches, such as sensors, opto parts and power packages. Larger competitors often are not interested, or simply not skilled, in providing these packages. However, as OSE USA's Dr. Fehr contends, "It's still the big guys that control the IC packaging business." Today new package technologies are coming, for the most part, directly from the packaging foundries-or the packaging foundries and their customers-and not just from the the semiconductor makers. Wafer-level packaging is a term that has jumped to the mainstream. To some in the business, it means a new competitive challenge; to others it means the likelihood of more revenue. The impact of wafer-level packaging will become very visible to the packaging foundries if the bumping process becomes more readily available as part of a normal process flow, according to Haitham Hamed, director of business development at Silicon Turnkey Solutions, Manteca, Calif. Currently, says Hamed, only one or two of the wafer fabs available for fabless companies offer such a service. And, he adds, "It is still expensive by comparison to other chip-and-wire techniques.
"Having said that," he adds, "more and more packaging foundries will try to leverage the benefits of WLP to attract more business." New wafer-level packaging techniques have not seen the light of day, unfortunately, due to current market conditions, he says. "Once market forces change," Hamed maintains, "I predict more companies will try to switch to such processes to capitalize on a niche and what will become a lucrative market segment." Kevin Dibelius, worldwide marketing director of STATS, Santa Clara, believes that the technology and services provided by packaging foundries are likely to evolve as the WLP technology gains acceptance.
WLP a Logical Move "The demand for higher package density, better performance and smaller footprints is driving flip-chip processes such as direct chip attach and wafer bumping. The next logical step for packaging foundries is to move further toward WLP," he says. About convergence, Dibelius notes that as package sizes continue to shrink, "the distinction between services offered at wafer fabs and IC assembly houses will blur somewhat on bumping and WLP. "Regardless, packaging foundries will continue to provide distinctive value-added in assembly areas, such as subsystem packaging, SOC and DCA. Even as WLP continues to grow and evolve in the future, the market demand for wire bond technology will still be with us for quite some time." Will WLP land in the front end or backend? "It's difficult to say," reports Ed Combs, executive vice president of ASAT Ltd., Fremont, Calif. "At the last MEPTEC meeting, we noticed that a convergence of wafer-level packaging between wafer fabs and packaging foundries is becoming apparent. Ultimately, though, I believe that packaging foundries will be doing most of the wafer-level packaging."
|
Copyright © 2002