The semiconductor industry's past two cyclical slumps were not kind to Texas Instruments Inc. (TXN ) The Dallas company flubbed its planning for the ensuing recoveries--and paid for it with market-share losses. When sales plunged in 1996, TI was preoccupied with finding a leader to replace then-CEO Jerry R. Junkins, who died unexpectedly from a heart attack. Two years later, when demand waned again, TI was in the midst of divesting the bulk of its sprawling empire, including memory chips and defense electronics, so it could focus on products geared to the Internet.
Last year, however, when the industry slid into its steepest decline ever, TI played a much stronger hand. Sensing an opportunity to one-up its rivals, it tapped the $5 billion-plus in its coffers to buy new equipment and keep research and development humming. Now, it's revving up state-of-the-art production lines to spew out scores of new chips, which could bolster its leadership in digital signal processors (DSPs) and analog chips. Both are vital for products that handle real-world signals, such as digital cameras, MP3 players, voice-activated toys, and other multimedia gadgets. Of all U.S. chip companies, "TI has done the best job of managing this downturn," says UBS Warburg analyst Thomas A. Thornhill III.
There are signs that the gamble is paying off. When TI announces second-quarter results on July 22, analysts are expecting 9.5% growth in revenues, plus $104 million in income--its first quarterly profit in more than a year. Meanwhile, sales of TI's chips have surged 10% since the beginning of the year, double that of the overall industry.
Yet with demand for most high-tech gear languishing in the doldrums, TI could do everything right and still come up short. Just three months ago, chipmaker stock prices soared when execs began talking about a rebound on the horizon. It turned out to be a mirage. Many chip stocks are now testing 12-month lows. On July 3, Advanced Micro Devices Inc. (AMD ) cut its second-quarter sales projections for the second time in two weeks. More recently, suppliers of chipmaking equipment have also gotten slammed by Wall Street.
So what's buoying TI's sales? Skeptics say it's little more than a temporary uptick in chips for wireless phones as handset manufacturers recover from last year's inventory glut. That's lifting spirits at Qualcomm Inc. (QCOM ) and National Semiconductor Corp. (NSM ) as well. But since wireless-phone chips account for only about 10% of industry sales, and most other markets are stagnant, total semiconductor revenues will rise only 3% or less this year, barely topping $140 billion, according to market researchers. Says Needham & Co. analyst Dan K. Scovel: "This is not a year for recovery. This is a year for convalescence."
TI's fans, on the other hand, perceive a more fundamental shift. Merrill Lynch & Co. analyst Joseph A. Osha predicts that TI's cash position will swell by $2 billion over the next 18 months. Among the world's top 35 chipmakers, Osha sees only Intel Corp. (INTC ) throwing off more cash. Boasts TI CEO Thomas J. Engibous: "We've turned the corner."
The turnaround took some doing. Despite its cash hoard, TI wasn't beating its chest when the downturn began--it was beating down costs. In 2001, three aging plants were closed and 2,500 workers laid off. By spring, TI had slashed sales and administrative expenses by 34%. But certain budget areas were shielded from the austerity drive. One was a new factory on TI's Dallas campus that had been put on hold when business went south before. Even while cutting capital spending by 35%, TI continued plowing money into the Dallas wafer-fabrication plant, or fab--a total of $4.5 billion to finish it and revamp several other facilities.
Those investments could reap huge dividends, because "companies that spend money in the downturn gain market share in the upturn," says William J. McClean, president of market-watcher IC Insights Inc. in Scottsdale, Ariz. Last year, Intel CEO Craig R. Barrett made the same point. Successful companies, he said, don't save their way out of recessions--they spend their way out. European chipmaker STMicroelectronics employed that strategy and jumped from No. 14 in 1995 to the No. 3 spot worldwide last year.
For its comeback, TI is counting on more than just new capacity. In the second quarter, the Dallas fab shipped its first cell-phone DSPs produced on 12-inch silicon wafers. Compared with today's eight-inch wafers, the larger platters slash per-chip manufacturing costs 30% to 40%, because twice as many chips can be crammed onto each wafer. That means TI can undercut the prices of rivals using eight-inch wafers. Moreover, the new fab can print the smallest transistors yet--a mere 130 nanometers wide, or 700 times thinner than a human hair. These ultratiny transistors reduce power consumption, a major selling point among makers of mobile phones and other handheld gadgets. Finally, TI is switching from aluminum to copper wires so signals can zip through chips faster, boosting performance.
TI is hardly the only chipmaker adopting these new technologies. But Engibous asserts that only an elite group is tackling all three, including IBM (IBM ), Intel (INTC ), Taiwan Semiconductor Manufacturing (TSM ), and United Microelectronics (UMC ). DSP expert William I. Strauss, president of consultant Forward Concepts Co. in Scottsdale, Ariz., credits TI with grabbing a big head start over its chief rivals: Agere Systems, Analog Devices, and Motorola. What's more, Engibous claims, "a significant number of traditional semiconductor strongholds have opted out" of one or more of these technology upgrades--in particular, Japan's cash-strapped producers.
That's news to the Japanese. Even Roger C. Mathus, a former TI manager and now head of the Semiconductor Industry Assn.'s Asian Office in Tokyo, says Japan's chipmakers are not about to throw in the towel, despite huge capital-spending cuts in recent years. To offset skimpy individual budgets, Japanese companies are banding together in a new round of joint ventures reminiscent of those in the 1970s and '80s, which helped Japan drive many U.S. chipmakers to their knees. The consortiums working on next-generation chip technologies include most of Japan's powerhouses, from Fujitsu (FJTSY ) and Hitachi (HIT ) to NEC (NIPNY ) and Toshiba (TOSBF ). But TI COO Richard K. Templeton doesn't see a threat: "Technology development is not a team sport," he says.
Whatever happens across the Pacific, TI needs new products that tickle customers and stimulate demand. That's the mission behind its $1.5 billion R&D operation, which has produced some 500 new high-performance analog chips to be released this year. They'll be used in digital audio receivers, cable modems, other broadband-network hardware, and handheld computers. Together they should generate more than $1 billion in sales over the next five years. Meanwhile, TI's hot new DSP chips for next-generation cell phones are fetching twice the price of their predecessors. The company believes they'll account for 50% of wireless revenues by yearend, up from 15% today--and lift gross margins above 50%, vs. the current 35%.
The main short-term concern is the market for chips in wireless handsets, which represent 20% of TI's chip revenues. Engibous is hoping the recovery there has legs. But even if it doesn't, he paints a bright long-term picture. TI's trove of new DSP and analog chips are just what the doctor ordered for anything with images and sound. "These are what will drive electronic wealth over the next decade," Engibous says.
That vision is shared by TI's competitors. Last month, three biggies--Agere Systems (AGR.A ), Motorola (MOT ), and Germany's Infineon Technologies (IFX ), which has a 12-in. fab--announced plans to collaborate on DSP designs and manufacturing. And with PC growth stalled, Intel also wants to cut in on TI's dominant market shares: 44% of DSPs and 60% of wireless phones. "TI's position is weakening," says Raymond A. Burgess, Motorola Inc.'s chief of semiconductor strategy.
Clearly, DSP and analog chips, which are projected to outpace the overall market in coming years, are silicon's hot new poker game. The players are plunking down billion-dollar bets, and the winners may be those few companies with very strong stomachs.
By Andrew Park in Dallas, with Irene M. Kunii in Tokyo