QCOM--Real Men Rent Fabs (Forbes)
By Eric Nee
The building that houses Qualcomm's other great growth engine--its chip operations--is five miles from Jacobs' office. One of 30 Qualcomm buildings that dot the La Jolla area, it's down the block from Nissan's Design Center, where new cars are dreamed up, and across from a large facility belonging to SAIC, San Diego's homegrown defense-consulting giant.
Qualcomm's building is spare, with a rudimentary lobby--clearly a place meant for working engineers. It's a lot like Donald Schrock, president of Qualcomm CDMA Technologies. He's a no-nonsense guy, totally focused on building CDMA chips. Except that Qualcomm doesn't literally build them. Instead, Qualcomm is the world's second-largest "fabless" semiconductor manufacturer, on track to become the largest this year. Rather than spend billions of dollars to build its own chip plants, Qualcomm designs the chips and contracts with others, principally IBM, to make them. That's different from Intel, Motorola, TI, and other chip heavyweights. The fabless approach lets Schrock concentrate the company's money, energy, and 800 engineers on designing new chips, not making them.
Being fabless has paid off. Qualcomm sells 90% of the chips that go into CDMA phones. In the fiscal year ended September 1999, that translated into $1.1 billion in sales, more than one-quarter of Qualcomm's total revenues. Analysts estimate that Qualcomm will pass $2 billion in chip sales next year, with a pretax profit margin of more than 40%. Qualcomm has actually run a few competitors out of the business. Oki Electric, LG Electronics, and Sony all made CDMA chips for their phones but dropped those efforts in favor of buying chips from Qualcomm. Why? Because Qualcomm has been able to add features, like geographic-positioning circuits and MP3 music-playing capability, sooner than anyone else. What's more, Qualcomm's "chip sets" use fewer chips than those of competitors, making possible cell phones that are more compact and that need less power. The chip sets also cost less to produce, giving Qualcomm the freedom to cut prices or make higher margins.
Qualcomm's chip success stands in marked contrast to its failed efforts in other CDMA hardware. In the early years, when it was the only company backing CDMA, it had to supply everything a cell-phone carrier needed to provide the service. That meant building phones and infrastructure equipment as well as chips. For a while, Qualcomm looked a lot like such vertically integrated vendors as Nokia, Ericsson, and Motorola, except for one key thing--it was much smaller. Ironically, its very success at getting the industry to adopt CDMA crippled the handset and infrastructure business: As larger manufacturers entered, Qualcomm was unable to compete. Its infrastructure gear--antennas, switches, base stations, and the like--never garnered more than a few percentage points of market share against the likes of Nortel and Lucent. Qualcomm sold that business to Ericsson last year as a sweetener in the deal resolving the 3G standards. The handset business, which Qualcomm developed with Sony, sold a lot of phones but continually lost money. Qualcomm sold that business to the Japanese electronics firm Kyocera last year.
Schrock was happy to see Jacobs unload these dogs. Not only did the losses they generated put a drag on Qualcomm's stock, but the businesses also competed with the very manufacturers he was trying to sell chips to. "Competitors tried to use that" to sell against Qualcomm, says Schrock. Now, when a Qualcomm chip salesperson walks into Motorola, Ericsson, or Nokia, she doesn't have to explain why Qualcomm is competing with them.
Which is a good thing, because the CDMA chip business is about to get much tougher. Qual-comm has overshadowed smallish chipmakers like PrairieComm, LSI Logic, and DSP Communications. But last year Intel, the $29-billion-a-year king of the chip hill, spent $1.6 billion to buy DSP as part of a major push into communications semiconductors. And some observers expect TI, which dominates the markets for TDMA and GSM chip sets, to enter CDMA. Though most analysts think Qualcomm will fare well even with increased competition, few believe it can maintain its 90% market share. Nor does Qualcomm. "More than likely it's hard to hold 90% share," says Schrock. "We expect it to come down, but I'd be disappointed if it came down to 50%."
All the same, the wireless market is growing fast enough for Qualcomm's chip business to continue to expand. The company shipped some 50 million CDMA chip sets in 1999. Analyst Piecyk expects that to grow to 73 million this year and 100 million in 2001. And that's just CDMA chip sets. Some expect Qualcomm to shed its single-minded devotion to CDMA and introduce chips for other standards. Qualcomm's design expertise and portfolio of technologies, like geographic-locator circuits that work even inside buildings (acquired when it bought startup SnapTrack this year, for $1 billion), could be put to good use in these markets as well. |
