Texas Instruments' lunatic fringe

(FORTUNE Magazine) -- Gene Frantz has a big, impressive title: Texas Instruments (Charts) principal fellow and business development manager, digital signal processing. Which tells you nothing about how he spends much of his time: searching for and encouraging all manner of lunatics and visionaries.

Patient and relentless as a hunter, he stalks his quarry among TI's engineers as well as among academics, inventors and employees of small tech companies all over the world. "What I look for in these companies is the wild-eyed optimist who's going to tackle the market," Frantz says.

For example, in February he invited Israeli entrepreneur Josef Segman to show an invention of his to engineers at TI's annual developers' conference. Segman's device is a cousin to Dr. McCoy's tricorder from Star Trek: Point it at a patient, and it tells you his vital signs. It's not quite up to 23rd-century design specs yet - its blood pressure and other readings on the TI executives who crowded into a Dallas hotel room to check it out were off a bit - but the thing is on its way to working.

Frantz told Segman, "I see you've done this. But I don't believe what I see."

Frantz is the dean of an informal and amorphous group of TI engineers (and their peers and contacts outside the company) who call themselves the Lunatic Fringe. They are senior people who have been given free rein to follow their curiosity wherever it goes.

"There's this continuum between total chaos and total order," Frantz explains. "About 95% of the people in TI are total order, and I thank God for them every day, because they create the products that allow me to spend money. I'm down here in total chaos, that total chaos of innovation. As a company we recognize the difference between those two and encourage both to occur."

The spirit of the Lunatics - look everywhere for good ideas first, worry about turning them into products later - has suffused TI for years, begotten some of the company's greatest hits, and largely explains the success it is currently enjoying.

What TI mainly makes are chips. Not the digital microprocessor computer-brain kind that Intel makes, but practically every other flavor. Key among them are analog- and digital-signal-processing (DSP) chips, which turn analog information into digital information. Both are essential for almost every electronic tool or toy. Open an Apple iPod, a Nokia cellphone, a Dell laptop, or a Samsung HDTV, and you'll see chips with the TI logo.

The company does put its name where the public can see it on a number of products - notably its long-running line of handheld calculators - but many more of its innovations can be found in the guts of electronic devices that carry other people's brands.

CEO Richard Templeton, a 26-year veteran, says, "We have the ability to work with the biggest companies in the world and the smallest." As they say in Texas, it ain't bragging if you can do it. TI has quietly become the world's third-largest semiconductor company, behind Intel and Samsung, with sales last year of $13.39 billion and net income of $2.3 billion. It has roughly 12% of the $30 billion global analog chip market.

In each of the past four years, TI has grown revenue faster than the semiconductor industry average. Since 2003, TI's stock has moved up from below $15 to its current range around $30. A key to that success is the Way of the Lunatic.

There isn't anything as formal as a Lunatic Fringe membership list. "There are more Lunatics within TI than show up and raise their hands," says Frantz, who somewhere in his 32 years at the company (he can't recall quite when) began bestowing the appellation admiringly on colleagues who were willing to explore crazy ideas. How many are there? "Within TI, maybe no more than 100," he says.

The closest thing they have to regular gatherings is the weekly Sea of Ideas meeting, named after a call to action made by Templeton. ("There's a big sea of ideas out there in the world," he exhorted his engineers, "and I want you to cast a big net.") And there are plenty of people within TI who don't necessarily consider themselves card-carrying Lunatics but nonetheless partake of the Fringe's freewheeling do-it-yourself ethos.

Dennis Buss, a TI fellow and vice president, recalls a group of engineers recognizing a looming problem a few years back. Next-generation mobile phones, they realized, would be used for receiving digital TV, taking digital photos, playing 3-D games, and other battery-draining tasks. So in 2003 they set a goal for themselves: a power-management chip that could cut consumption 1,000-fold. And a deadline: the end of 2004.

"No marketing people, no businesspeople," Buss says. "For almost two years there were meetings at 7 A.M. Guys in Japan were giving up their Friday nights. People in India and Nice were presenting data. This was not anybody's job. We were spending our own time to get things going."

They made their goal - with five hours to spare - and the resulting SmartReflex technology is now partly responsible for TI's success in the mobile-phone market. Buss says that at any time there may be a dozen or more such ad hoc groups working inside the company on under-the-radar projects.

What Buss is describing is Lunatic Lesson No. 1: Good things happen to the bottom line when engineers pursue projects because they are jazzed by a tough technical challenge, not by a mandate from the marketing department.

Born in the oil patches of Depression-era Texas and now employing 35,000 people around the world, TI developed the world's first transistor radio. A TI engineer, Jack St. Clair Kilby, created the first integrated circuit and won a Nobel Prize. Millions of people made it through high school math clutching handheld calculators, another Kilby creation.

During the 1980s, though, TI seemed to lose its focus on innovation. As the digital-chip market exploded, TI was getting its head beaten in by competitors, beginning with Intel. The company had also saddled itself with a range of underperforming businesses like defense, PCs, and memory chips and had copped an arrogant big-company attitude.

But by the early 1990s, the tech world had begun shifting toward a range of ever smaller devices - laptops, cellphones, portable music players - that required the analog and digital-signal-processing chips TI had pioneered.

"A funny thing happened on the way to the digital revolution," Art George, head of TI's high-performance analog business says. "There's a perception that as things become more digital, the need for analog goes away, and it's completely the opposite."

Under the leadership of Jerry Junkins, then TI's chairman, president, and CEO, the company began reinventing itself. After Junkins died of a heart attack on a trip to Germany in 1996, Tom Engibous took over and shed businesses that didn't fit, including PCs and defense. Engibous also decided to focus on DSP. That technology, which now accounts for 40% of TI's revenue, first appeared in the quintessential TI Lunatic Fringe project.

In the 1970s, Frantz started working on an idea that eventually became Speak & Spell, the talking educational toy. "I started out doing Speak & Spell under the table," Frantz says. "It wasn't an official program until we'd had it running about six or eight months. We told management at that point, and they said, 'Okay, we can't kill it. What's it going to take to finish it?' "

Speak & Spell was a hit, but more important, it contained TI's first DSP chip. At first, turning analog signals like sound waves into digital signals a computing device could understand or transmit was just an interesting engineering problem. When cellphones and other portable devices came along, it became a huge and very profitable business.

During its wilderness years, TI fell into the bad big-company habit of ignoring little customers. Nowadays the Lunatics and their like-minded colleagues spend a lot of time traveling to meet entrepreneurs in tiny companies on the theory--proven by their own experience going back to Speak & Spell - that today's wacky idea can turn out to be tomorrow's billion-dollar industry.

"Those four customers in the garage might be the next Apple or Sony or the next Dell," says analog-chip chief George.

When Craig Malloy and Michael Kenoyer started their company, LifeSize Communications, it had just two employees - them - and an idea for bringing videoconferencing into the high-definition era. Malloy had talked to TI a decade earlier at a previous startup. TI barely gave him the time of day, and he went with a processor from Philips.

This time, Malloy says, TI "brought over eight or ten people and spent three or four hours talking to us and asking us what we wanted."

The result is LifeSize Room, an $8,000 to $12,000 system that delivers high- definition video and CD-quality sound over an Internet connection. It uses five TI DSP chips and 70 other TI components, along with custom silicon and chips from TI rivals. LifeSize says it has sold 1,000 systems - the latest to a casting agency that allows actors in New York City and Los Angeles to audition for roles on the opposite coast via teleconference.

TI has also had a hand in one of the coolest new products this year via another small customer, Sling Media. Its Sling Box allows users to "place shift" video from their home TVs and video recorders to a laptop or cellphone anywhere in the world over the Internet. Co-founder Bhupen Shah says TI started working with Sling when his company had four employees.

Think of this virtuous cycle - curious, open-minded engineers finding pockets of innovation around the world, which in turn inspires further innovation back at HQ - as an extremely effective opportunity-detection system. The intersection of entertainment and communications, where Sling and LifeSize reside, is one area the company is counting on for growth. Health care is another. A personal passion of head Lunatic Frantz, for instance, has been a project at the University of California's Doheny Eye Institute to develop a retinal implant that would allow blind people to regain some of their vision. (TI chips are also in prototype bionic limbs currently being tested in Chicago.)

TI has reasoned that it pays to think small, to work with tiny customers or with academic researchers. A successful idea could scale very quickly if it works. That small-is-beautiful logic makes particular sense in the health field, where startups are routinely absorbed by established players.

According to global strategic marketing VP Doug Rasor, "Given the rhythms of the FDA and how long it takes to get a product developed, tested, submitted, retested, and approved, the big guys get their new products from the acquisition of small guys And by the time they get acquired, you can't come in after the fact and say, 'Will you redesign this part of your system because we want to be your friend?' "

TI, of course, has to balance its interest in little guys with the needs of what its sales reps call HICs (huge important customers) like Nokia (Charts) and Dell (Charts). TI power-management and hard-drive controller chips are used in three of every four notebook PCs. Seven of the top ten digital-camera manufacturers are customers. A digital light-processing (DLP) chip developed by TI is the core of millions of high-definition TVs and 99% of new digital movie-theater projectors.

The question, as it always is for a company so dependent on innovation, is, What comes next? TI has high hopes for a device that combines the functions of a cellphone on one chip and for its new DaVinci DSP chip, a complex system of silicon, software, and development tools that can be "tuned" for any video application.

CEO Templeton believes DaVinci will be huge. "I do have a strong belief that video is going to really matter, just not on the timetable where people want to fill holes in the third quarter of next year's spreadsheet," he says. "I put this in the context of where we were in 1985 in the DSP business. We would not have sat around at that point and predicted what some of those products were going to look like."

There is always the chance that TI could lose touch with the Way of the Lunatic. It's happened before. Gene Frantz doesn't seem too worried. He points out that TI's current bets show no signs of busting anytime soon.

And Frantz and his fellow Lunatics are still happily on the hunt for the next crazy idea. "The ones that excite me," he says, "are the ones we don't know about yet."