MUST READ
pathfinder.com
Hotter Than Intel
Part 2
Erick Schonfeld
Programmability also helps keep DSPs cheap. Since the major difference between, say, a DSP in a cell phone and a DSP in a digital camera is the software that tells it what to do, startups developing products for new niches can piggyback on the manufacturing efficiency that TI gains by supplying high-volume customers like cell-phone giant Nokia. And whereas a few years ago TI was a laggard in chipmaking technology, today it is a world leader. It makes a cell-phone chip whose circuits are a mere 0.18 micron (millionth of a meter) apart, the same as in Intel's and IBM's most advanced chips.
DSPs and analog chips make up about half of TI's sales; much of the rest derives from other chips, such as the processors that run Sun Microsystems' workstations and servers. Thanks to Engibous' aggressive buying and selling of divisions, operating margins have expanded from about 4% to 19% in the past three years, and TI's business portfolio has substantially changed. Gone are the cyclical military-contracting and memory-chip units; the only non-chip divisions TI still owns are calculators and electrical controls. At least one of these deals, the sale last year of TI's money-losing memory-chip business to Micron Technology, has the makings of a financial coup. TI received 41 million shares of Micron stock (partly in options), currently worth about $3 billion. Says CFO Bill Ayelsworth, "It is our intention to sell those shares in the next memory-chip up cycle." Engibous has hinted to analysts that TI would like to cash out when Micron shares hit $150, double where they trade today. At that price, the sale would mean a $6 billion bonanza.
What would TI do with so much cash? For one thing, it could continue snapping up hot DSP and analog startups. In the past three years, it has spent about $2.5 billion to buy companies with expertise in high-speed ADSL modems (Amati Communications), cable modems (Libit Signal Processing), Internet telephony (Telogy Networks), wireless networking (Butterfly VLSI), and analog chips (Unitrode).
While TI's DSPs and analog chips are expected to go into at least 150 million cell phones this year, it faces stiff competition. Qualcomm, notably, has a virtual lock on DSP chips for cell phones based on the CDMA standard, which is growing faster than any other. Next-generation phones that will be able to send e-mail and surf the Net will probably use a version of CDMA. But TI is on the ball. Its DSPs are being designed into eight of the top ten next-generation cellular base stations. And in the race to supply chips for modems and disk drives, TI must contend with Analog Devices, Broadcom, IBM, and Lucent Technologies.
To meet Wall Street's lofty expectations, TI also must find growth in markets that have yet to take off. To do so, the company tries to spot markets just as they are beginning to gather steam. It helps engineering professors at more than 900 universities teach students to program its DSPs. An estimated 30,000 students have taken the classes in the past 15 years--an asset no other company can duplicate. And by making available a broad catalog of standard DSP and analog chips aimed at the niche companies where most of these programmers work, TI assures itself access to some of the most creative engineers on the planet.
This catalog business, the second-fastest-growing part of the company after cellular, is a key area of focus and inspiration. About a year ago, Greg Delagi, who runs the unit, noticed that TI's off-the-shelf DSPs were being designed into a new type of consumer-electronics product: digital audio players that can download music from the Internet. "At first, we had no idea what an Internet audio player was," he admits. "Then we started winning one design after another, and all of a sudden it became a cluster." Now TI has an entire business unit devoted to seeding this nascent market with its DSPs. By Christmas, digital music players powered by TI chips will be sold by RCA, Sharp, and even Lucent, whose own chip unit is the world's No. 2 maker of DSPs.
Says Doug Rasor, TI's head of emerging DSP businesses: "We really do believe DSPs will show up in all kinds of places that are hard to predict. So we are trying to cast the widest net possible." In addition to Internet audio, TI now has minidivisions for digital cameras, printers, Internet telephony, digital speakers, handheld information appliances, electric-motor controls, and wireless networking equipment. Rasor's goal is to start three to four such units each year; he identifies niches by analyzing the patterns of the catalog shoppers, and looking for those who are making things with the potential to be the next cell phone or modem. Says TI president Rich Templeton: "These people find us on the Web. The catalog market is sheer numbers and brilliance working for us. Later, we paint a bull's eye around the dart and call it a strategy."
TI's insight into these potentially huge digital markets has investors salivating. By serving thousands of customers in the low-volume catalog market, the company increases its odds of being a primary supplier to whoever the next Nokia may be. Up till now its execution has been flawless, but juggling so many new projects has risks. Unless TI can tend all the pots on the stove, it may end up serving tasteless gruel to investors instead of a savory feast.
The man at TI with the deepest perspective on these opportunities is Jack Kilby, one of the two men who invented the integrated circuit, in 1958 (the other was Robert Noyce). He comes in about once a week to the Kilby Center, TI's gleaming black R&D facility in North Dallas. His spare, windowless office--furnished with little more than an extra-high table and chair to accommodate his 75-year-old, 6-foot 6-inch frame--contrasts starkly with the astounding proliferation of products based on his work. Without Kilby's invention, there would be no pocket calculators, digital watches, PCs, modems, PalmPilots, or cell phones. There would be no Internet, no information age. Last year he reflected in an interview with FORTUNE: "The industry has always grown by finding new high-volume applications. The calculator, watch, PC, cell phone--each was an order of magnitude higher than the last." He paused to chew over this fact, then added, "The fascinating part is that it is not over." Not by a long shot. |
