Just in case this article hasn't been posted (sorry if it has) ..seen on MF
MESSING WITH TEXAS
CEO Tom Engibous tells how Texas Instruments is transforming itself again - this time into a digital signal processor powerhouse.
By Nikki Goth Itoi
In mid-1996 Texas Instruments appeared to be in trouble. The company's respected CEO of 11 years, Jerry R. Junkins, had suffered a sudden and fatal heart attack; DRAM prices, which still generate about 12 percent of TI's revenues, were plummeting; and the company was in the earliest stages of a major transition that would increase its emphasis on the digital signal processor (DSP) business. By July of that year, TI's stock (NYSE: TXN) had fallen from a high of $39.94 the previous September to a low of $21.63. For Tom Engibous, the former president of TI's semiconductor group and Junkins's successor as CEO, it was an inauspicious beginning.
Since then, however, Mr. Engibous and the other members of TI's senior management have won over even the company's harshest critics by not only sticking to the initial transition plan but accelerating it rapidly. DSPs and mixed-signal processors now bring in the largest chunk of TI's business, 32 percent. And TI has aggressively divested nine business units, including notebook computers and defense electronics, and acquired six new DSP companies: Tartan, Silicon Systems, Intersect Technologies, Amati Communications, Go DSP, and Spectron Microsystems.
DSPs are powerful microprocessor-like chips that can handle digital signals in real time. They are intended for communications applications that can't tolerate delays, like cellular phones and satellite TV broadcasts. Dataquest values the DSP market at $3.3 billion. And with an annual growth rate of 30 percent for revenues and 37 percent for unit production, DSPs are the fastest-growing segment of the semiconductor industry, according to the market research firm In-Stat.
Although Lucent Technologies, which invented the DSP, has been the dominant DSP vendor for telecommunications applications, TI has been the leading DSP supplier overall since 1982, when it introduced the first commercially available DSP. In-Stat reports that TI today holds a market share of 40 percent--slightly greater than that of its two closest competitors, Lucent (28 percent) and Analog Devices (11 percent), combined. Mr. Engibous believes his company can widen that lead, even as competition increases, by continuing to narrow its focus.
Though TI has seen a 30 percent decline in total revenues in the past 18 months, Wall Street is pleased. The company has doubled its market value, with shares trading in the low $50s in February. And it doesn't appear as if TI's restructuring is finished yet. Although Mr. Engibous declines to comment on the subject, rumors have recently surfaced that TI will start phasing out its DRAM business next.
Nevertheless, analysts question whether TI will be able to hold its own against new competitors that are already folding the DSP architecture into microprocessors and microcontrollers to produce an even more compelling chip. Just after TI announced that its 1997 year-end revenues were up 10 percent (against total semiconductor market growth of only 5 percent), The Red Herring asked Mr. Engibous to explain TI's remarkable recovery and to expand on the company's future strategy.
The Herring: At what point did you realize the DSP market offered a significant opportunity for TI?
Engibous: We knew it was a good business back in the early '90s. In 1992 and '93, we knew we wanted to pull out the stops and see how big this thing could be. But it remained only one segment of a large, multifaceted company. Finally, in 1996, we recognized that the DSP market could provide the growth, value, and stability that we wanted--but quite frankly had not been able to achieve in the past. So we made the decision then to focus our resources very heavily on DSPs.
The Herring: What is the key to TI's phenomenal growth rate in the DSP business?
Engibous: To say we've done everything perfectly is far from the truth. We introduced our first DSP in 1982. And whether that was brilliant foresight or a good bet on a potential opportunity--which is probably far more accurate--the reality is that it's now a large market, and we find ourselves in a dominant position.
The key is probably that, first, we've tried to act more like a startup than a large, 68-year-old company. That's easy to say but not easy to do. There aren't many $10 billion companies that are focused as narrowly on something as we are now. Second, with our leading position, we are able to spend as much on R&D as the revenues of our nearest competitors. That doesn't preclude other companies from entering this market, but it means we are able to capture a vast amount of the growth. And third, we're the only DSP company that has applied leading-edge microprocessor-type silicon technology to DSPs. We can run DSPs with clock rates similar to those of Intel or IBM, and our competitors don't have access to that technology.
The Herring: Tell us about the most exciting future applications of DSPs.
Engibous: The move to digital communications is going to play a crucial role in determining the future applications of DSPs. Today, to my knowledge, every digital cellular phone and almost all high-performance modems involve a programmable DSP. Wireless communications and Internet access features are going to be important in the near term because they require compression and modulation, both of which are performed best by DSPs.
SOUTHBOUND Texas Instruments' DSP acquisitions Target Date Value ($M)
------------------------------------------------------------------------
TARTAN 5/1/96 N/A
------------------------------------------------------------------------
SILICON SYSTEMS 7/1/96 575
------------------------------------------------------------------------
INTERSECT TECHNOLOGIES 3/1/97 N/A
------------------------------------------------------------------------
AMATI COMMUNICATIONS 11/1/97 395
------------------------------------------------------------------------
GO DSP 12/1/97 N/A
------------------------------------------------------------------------
SPECTRON MICROSYSTEMS 1/1/98 20-26
------------------------------------------------------------------------
However, the largest segment of our DSP business today is not wireless or mass storage but what we call mass-market applications. Perhaps the biggest space within this broad category is video consumer electronics. The future of televisions, VCRs, camcorders, set-top boxes, and still photography is digital. And in every single one of those devices, the data is compressed to some format and then decompressed. And those are DSP functions.
The other area that's interesting is digital control of automobile motors. Today very few motors are digitally controlled, but the opportunity for increasing energy efficiency, lengthening the life of the motor, and ultimately lowering the cost of the car indicates that, with time, all motors will be controlled this way.
The Herring: Because this is a newer market, DSP tools (the software needed for programming the processors) are far less advanced than those available for microcontrollers and microprocessors. Is TI working to change that?
Engibous: DSP tools are still somewhat primitive--not only in terms of ease of use, but in how the tools work together. Presently, DSP tools are much more individualistic than microcontroller code, and the code development that goes with a DSP is far more intense and costly.
We've made several acquisitions in the past 18 months [see table]--and we might make more in the future--to address this problem. We would like to move DSP technology out of the hands of laboratory-type programmers and into the hands of mainstream programmers, so that people who are comfortable writing software for microcontrollers could also write software for DSPs.
The Herring: How much of TI's overall R&D budget is currently slated for DSP technology?
Engibous: We spend about $1.2 billion a year on R&D, and the largest segment of that goes into the DSP area, including silicon technology for the leading-edge DSPs. We're also spending outside TI to promote the DSP market. Last year we set up a $100 million venture capital fund and a $25 million university research fund for investing in DSP technology. We've already exceeded our expectations in terms of the number of startups now interested in communicating with TI, and we've committed to about six investments so far through the venture capital fund. We also help universities in the United States, Europe, and China establish courses in TI's DSP architecture.
The Herring: So has TI completely reinvented itself?
Engibous: This is actually TI's third reinvention. TI began as an oil-field service company. Then it became predominantly a defense electronics company. We licensed the transistor from Bell Labs in 1950, invented the integrated circuit in 1958, and subsequently became the largest semiconductor company in the world. And now we're well on our way to becoming a DSP company. Our DSP business has grown 30 to 40 percent a year for the last several years. We grew faster than the market in 1997, and in the fourth quarter we reported record revenues, profits, and orders in our DSP and mixed-signal products division. I expect that we'll grow 30 percent again in 1998.
In the past 18 months, we've allowed total revenues to drop by 30 percent, but in total value we're up at least twice from where we were. Even with the current memory downturn, we're running at the highest profit levels in our history. But we're not content merely to maintain that lead. We want to work aggressively to widen it. |
