Texas Instruments hit the comeback trail in the 1990s. Since assuming the top post in 1996, Tom Engibous has shepherded the electronics giant through a radical transformation, divesting its massive defense business and refocusing instead on digital signal processors and analog integrated circuits. Today, crack open any digital cell phone, and you’ll probably find a map of Texas-the TI logo. In an e-mail dialogue with technologyreview.com staff editor Alan Leo, Engibous discussed TI’s evolution, the state of digital signal processing and the future of wireless computing.TR: You’re the rare engineer among Fortune 500 CEOs. Compared to B-school execs, what advantages do you draw from your technical background?
ENGIBOUS: There’s a difference between an engineer who asks another engineer to do the seemingly impossible and that same request coming from somebody with a pure business background. The technical background helps build a camaraderie and confidence between the people who are handling the technology and the people who are running the business. It reduces the typical “us and them” mentality that often prevails between management and core workers.
TR: And, as an engineer, what disadvantages have you had to overcome?
ENGIBOUS: Perhaps I had a little steeper learning curve when it came to the non-technical aspects of running a Fortune 500 company. But I can say in all honesty, I wouldn’t have done it any other way. In fact, I don’t think I could have done it without an engineering background.
TR: What has been your most satisfying professional moment?
ENGIBOUS: My first analog circuit design.
ENGIBOUS: Because it worked, and I knew I had a future at TI.
TR: And more recently?
ENGIBOUS: More recently, though, it’s not about a satisfying “moment” as such, but the general success that TI has achieved through its focus on real-time signal processing via programmable DSP and analog technologies. In the past few years, we’ve gone from being viewed as a dinosaur to being recognized as one of the world’s most dynamic technology companies. That’s been very satisfying.
TR: What changed?
ENGIBOUS: DSPs remained a niche business for us until the mid-1990s, when it became apparent that real-time signal processing via DSP and analog semiconductors was the right capability at the right time. So, we evolved again and became a DSP and analog company. This has dramatically narrowed our focus, but it has greatly expanded our business opportunities. But if you think about it, we’re still processing signals, which is what we did in 1930. The difference is that now we do it in real time.
TR: So what was your worst moment?
ENGIBOUS: Probably when we divested the defense business in the mid-1990s.
ENGIBOUS: It affected many people who had been part of the TI community for a very long time. The defense area was an important part of TI’s origins, and it led to significant innovations through the years. Making decisions that have a negative impact on people-even if it’s for an ultimately greater good and benefits the vast majority of employees-is probably the worst part of this job.
TR: How did you handle that?
ENGIBOUS: We started by honestly and carefully explaining to employees our objective, which is to lead the world’s evolution to real-time signal processing. To accomplish that, since 1996, TI has made 22 divestitures involving more than 23,000 employees, and 23 acquisitions of companies that employed over 6,000 people in total. That can be traumatic at both a personal and corporate level, so we’ve tried very hard to treat individuals with genuine respect and help them make a successful transition whether they remained within TI or not. I don’t know that we’ve always succeeded at that-probably not-but we’ve genuinely tried.
TR: How did TI preserve its ‘core culture’ during all this change?
ENGIBOUS: Reinventing the company as markets evolve has always been part of TI’s “core culture.” Today, we’re the world’s leading DSP and analog semiconductor company. But we started out in 1930 as an independent reflection seismography contractor for oil exploration. By 1940, we’d grown from 12 crews in Texas “wildcat” oil patches to become an international company working on three continents. Building on that expertise, we extended into submarine detection in the early 1940s and airborne radar detection in the late 1940s. By the time we introduced the first single-chip Digital Signal Processor in 1982, oil exploration was part of TI’s distant past, so change is nothing new to us.
The Next Next Big Thing
TR: You have said that TI hopes to do for the mobile Internet what the Web browser did for the immobile one. Where do you see the “killer app” for wireless Internet?
ENGIBOUS: I believe that adding mobility is the next major step in the Internet’s evolution, just as the invention of a graphic Web browser was the match that ignited an explosion in usage of the fixed Internet. Will the killer wireless app be global positioning satellite and location-based services? Speech recognition? Audio and video streaming? Mobile commerce? The truth is, I don’t know. And a bigger truth is, nobody else knows for certain either.
I do know that if the mobile Internet is to be successful-if it’s going to live up to its potential to add new dimensions to Internet usage-then we have to get the fundamental building blocks in place. We need the infrastructure, which is being installed right now. We need the portable devices, which are in development and early production right now. The killer app doesn’t matter if there’s no infrastructure to transmit it and no handsets that can use it. I’m glad to report that these building blocks are continuing to come into place rather well, despite the current economic slowdown.
TR: What are the biggest obstacles to ubiquitous computing? How is TI tackling them?
ENGIBOUS: First, I think we need to change the term a bit. “Ubiquitous computing” tends to imply a mobile World Wide Web, or working on a financial spreadsheet while you’re sitting in the park. If that’s what an individual user wants, fine, but I think it goes beyond that.
TR: What term would you use, then?
ENGIBOUS: What TI is working toward is real-time wireless multimedia and communications-real-time signal processing that encompasses computing, adds new dimensions and delivers it anytime, anywhere. That experience begins and ends with communications, and what happens in the middle is content-information and entertainment. Of course, it’s a lot harder to say “real-time wireless multimedia and communications” than it is to say “ubiquitous computing.” But the former more accurately describes what’s really evolving before our eyes.
TR: In that case, what’s the biggest obstacle to real-time wireless multimedia and communications?
ENGIBOUS: The biggest obstacle is dealing with the vast complexity of high-speed wireless, and doing it in a very energy-efficient manner. We also have to accommodate diverse network protocols such as GSM [Global System for Mobile Communications], WCDMA [Wideband Code Division Multiple Access] and others. We have to support multiple operating systems such as EPOC, PocketPC, Palm and Linux. We have to function within standalone and combined form factors such as wireless phones, PDAs, laptop computers, Internet audio players and digital imaging devices. Security will be critical for achieving consumer acceptance, including firewalls, virus protection, encryption, digital signatures, biometrics and copyright and content protection.
We have to be backward-compatible to previous standards as well as delivering battery life that equals or exceeds what end-users are accustomed to today. We need to handle converged media, including voice, audio, video, images and data. We need to interoperate across multiple networks-wide area, local area, home and personal area. Shall I go on?
TR: Where do you see TI’s toughest competition coming from?
ENGIBOUS: The semiconductor industry is brutally competitive, but TI has achieved an enviable position, and our future is in our hands. I do have to concede that TI has no control over short-term macroeconomic conditions that affect consumer demand.
Some of the competitive alternatives being promoted now include field programmable gate arrays and RISC-centric architectures for wireless handsets. I think one of the reasons some companies promote a non-DSP approach is because TI has such a huge lead when it comes to programmable DSP and analog technologies.
TR: Where do you see the biggest opportunities for growth in your industry?
ENGIBOUS: The two biggest growth opportunities we see over the next few years are real-time wireless multimedia communications and delivering broadband to and throughout the home. Close on the heels of these markets, though, we see great things ahead for digital audio in general and Internet audio in particular, digital imaging, digital motor control and optical networking. I also think that digitized personal medical devices and services are a great area.