Unless there is a dramatic turn-around in the mobile phone business, the GaAs industry is headed for a downturn in 2001. This was the opinion offered by Dale Pfau of CIBC World Markets at the 2001 Key Conference, which was held in Key West, Florida on March 56. Pfau said that while GaAs device shipments have seen several years of steady, even spectacular growth, they may decline by as much as 15% in 2001, causing the first recession in the GaAs industry s brief history (see ). And while other presenters and attendees were of the opinion that the CIBC outlook was unnecessarily pessimistic, a rash of earnings warnings from major GaAs fabs, issued while the conference was in session, seem to support Pfau s conclusions. The slowdown is due to a sudden and sharp drop in the production of wireless handsets. Pfau told the audience that in February CIBC cut their forecast for 2001 handset sales from more than 500 million units to 475 million units. In comparison, approximately 410 million units were sold in 2000. But, Pfau noted, there is already a great deal of inventory in the pipeline, with perhaps as many as 50 million handsets having gone unsold in 2000. He told the audience it is therefore conceivable that production of mobile phones will show virtually no growth in 2001. In this set of circumstances, Pfau believes that handset manufacturers will exert pressure on all their suppliers, including the GaAs IC companies, to lower their average selling prices (ASPs) more quickly than would otherwise be expected. Under pressure It is this "double whammy" lower unit volumes and lower ASPs that leads Pfau to the conclusion that 2001 will be a disappointing year and that these woes may persist into 2002. "The GaAs industry, which has thrived on performance, is now going to be faced with price pressure and I think that is a significant threat," said Pfau. Eight months ago, in the pages of this magazine, CIBC set out a fairly rosy forecast for the GaAs industry (see Compound Semiconductor August 2000, p27). Not surprisingly, conference attendees wanted to know what lay behind the dramatic downgrade of the short-term outlook. Pfau explained that the collapse of the stock market is a significant factor. Though now more than a year old, the bear market was, until recently, blamed largely on the bursting of the "Internet bubble", with the assumption that "real" companies would soon bounce back. Now, however, the markets are denying new capital to telecom carriers and equipment companies alike, forcing them to curtail plans to build up more infrastructure and causing a major slowdown in efforts to attract new subscribers. This therefore results in far fewer new handsets being needed. Pfau also noted that there are virtually no new handsets on the market to generate consumer enthusiasm, and so the replacement market is also slowing down. New services that require new phones are also lacking. Pfau noted that wireless application protocol (WAP) services, which were introduced last year to provide mobile phone users with rudimentary Internet connections, have been a total disaster. And there have been significant delays in the roll-out of GPRS phones, the so-called 2.5G handsets that are supposed to provide a vast connectivity improvement over WAP phones. GPRS is now expected to be introduced in Europe in the second half of this year. Pfau advised the industry to keep a close eye on this development. "If GPRS is a success, it will provide the carriers with more revenue and should speed the transition to 3G, and another round of handset buying," he said. "However, if GPRS is a failure, then the recovery scenario for GaAs in 2002 is in jeopardy." Winners and losers Pfau predicted that the current turmoil in the handset market would create opportunities for companies offering RF modules as opposed to RFICs. Some GaAs companies, most notably Alpha and RF Micro Devices, are already making strides in this area. However, the move to modules may also allow silicon IC manufacturers to gain market share. "Modules have the potential to be disruptive because you are not required to have the best-of-breed technologies inside them," Pfau noted. "We have actually seen companies with a weak silicon-in-module format beat guys with good GaAs technology because they had good packaging or maybe unique filtering technology inside." And while Pfau believes that all of the established GaAs fabs have sufficiently large financial reserves to ride out this storm, he did state unequivocally that new companies offering GaAs foundry services were going to suffer. Several ventures of this type are in the works, >> especially in Taiwan. While the foundry model continues to be popular in the silicon industry, Pfau noted that the downturn he was describing means that there is no longer a lack of capacity in the GaAs market. "We don t see the need for any additional [GaAs] capacity for the next two years," said Pfau. "Under this scenario, most of the current GaAs foundry projects may not survive." When asked to identify some possible silver linings behind the gathering dark clouds, Pfau expressed enthusiasm for millimeter wave point-to-point radios but not other high-frequency applications such as LMDS and MMDS. Pfau also explained that he is very enthusiastic about the prospects for InP technology. "I m as bullish about InP as I have ever been about any sector that we ve looked at," he told the audience.