The first thing to understand about Navdeep Sooch is that he does not do things casually. There is always a purpose, a goal. When he decided to take up golf four years ago, it wasn’t because he wanted to drink beer with friends at the nineteenth hole. His objective was to qualify to play in the Buckman Cup, an elite annual tournament at Austin’s Barton Creek Resort and Club. So he took private lessons. He hit balls at the driving range four times a week. His score dropped precipitously, and that spring he not only qualified to play in the Buckman—a remarkable feat for a rank novice—but also won his flight in the club championship tournament. Sooch would never share such a story. Like most of the other interesting details about him, it comes from his wife. But this sort of determination is precisely what has pushed him and his company, Silicon Laboratories, to the forefront of one of the hottest industries in the world: specialty computer chips. When Sooch, now 38, co-founded Silicon Laboratories in 1996, he left a $149,000 job and took no salary for six months, betting everything he had on a new type of chip design. Today, as Silicon Labs’ chairman and CEO, his shares in the company, which designs chips that make modems and cell phones more efficient, are worth $200 million. Those shares were worth as much as $522 million last year—which landed Sooch on Fortune‘s “America’s Forty Richest Under 40” list. That was before the high-tech Armageddon in the stock market began in April. In spite of the market drop, industry analysts are confident of Silicon Labs’ future. And even at the lower valuation, it’s not a bad haul for three years’ work.

The second important thing to understand about Sooch is that he would rather do almost anything than say something revealing about himself. His stripped-down second-floor office in a nondescript building in southwest Austin tells you little about the man, other than that he is spartan, precise, and—this he will admit—”no nonsense.” He wears sensible black slacks and a beige cotton sweater. Small, neat stacks of manila folders line the top of his desk, where there is but one item that could be considered personal—a hand-painted coffee mug that his daughter, Kelly, made for him for Father’s Day when she was three. He never uses it, though, because he rarely drinks coffee. His recent award from Ernst and Young as the area’s entrepreneur of the year doesn’t hang on the wall. Neither does the framed stock certificate from when Silicon Labs went public, on March 24, 2000, and which quickly boosted his net worth above half a billion. They’re both at home, decorating his wife’s office. He refuses to let her order the plaques for the 24 patents he has received. “No self-respecting engineer would put those things up,” he says. “It’s too pretentious.”

His silver BMW 740i and red Porsche Carrera, however, are apparently not too pretentious. Cars are his one weakness, as it turns out, and it is comforting to know, at least, that he has one. He drives both cars at speeds that sometimes top 100 miles per hour along the curvy, hilly, two-lane road from his five-bedroom Westlake Hills home to his office. This detail does not come from Sooch either. It is related by his wife, who has heard the breathless accounts of their three children, who have been along on some of the joyrides.

When asked to describe what, exactly, he does, Sooch, leaning forward over his desk, is immediately evasive. “My role is . . .” he says, pausing for a moment and smiling coyly, “I’m the leader. So I don’t do anything.” This will no doubt delight his shareholders, who have already watched the price of their shares fall from a high of $105 in March of last year to a low of $10 in January. Of course, he’s lying. Silicon Labs is not a fly-by-night dot-com and didn’t grow from zero to $100 million in annual revenues in just four years’ time—up 100 percent over last year—with a chief executive who does nothing at all. Walking past an empty office on a Monday afternoon, Sooch hesitates for a moment, as if he really has forgotten that he’s the leader, and then slips in. He picks up a plastic container resting on the desk, where a single black chip, just three-eighths of an inch long, is nestled in gray foam. This is one of his products. “Here’s one,” he says and angles the chip toward the light, a bit of real enthusiasm creeping into his voice. “See where the plastic is etched off? That’s the actual silicon. It’s beautiful.”

Navdeep Sooch was born in Amritsar, in northern India, in 1962, the second child of Shangara and Pritam Sooch. Both had been raised on farms in the agrarian state of Punjab, and after their marriage they wanted to break free from the country’s rigid caste system. They went to college, then moved to Detroit. As it turned out, degrees from India didn’t help them find work in the U.S., so they both enrolled in Wayne State University, while Nav and his sister, Jasdeep, stayed behind in Amritsar, where they attended Catholic boarding school. A few years later, when Sooch was ten, he and his sister joined their parents. Living in a small apartment in an area known as the Cass Corridor, one of the city’s roughest neighborhoods, he became more fluent in English, picking up slang to go along with it. His mother worked as a high school librarian; his father was a systems analyst at Michigan Bell. They lived frugally, and within a few years the family had saved enough to buy a modest house in the suburbs.

Sooch attended the University of Michigan and majored in the closest thing to electronics that he could find—electrical engineering. He quickly learned the language of ones and zeros, but he found digital design unchallenging. He focused instead on the rarest and most difficult type of electronic circuitry design—so-called mixed-signal—which takes analog signals like sound and radio waves and converts them into finite digital numbers for computers.

After graduating at nineteen, Sooch worked for AT&T Bell Laboratories and went to graduate school in electrical engineering at Stanford University, where he sometimes skipped classes and borrowed notes from another electrical engineering student, Janet Harman. Within three years, he got his master’s degree, married Harman, and received four patents for integrated-circuit chips. In 1985 Sooch took a job with the Austin start-up Crystal Semiconductor because he was so impressed with Dave Welland, an MIT graduate who, instead of conducting an interview, argued with him for hours over the finer points of mixed-signal design. They also found a common bond in their favorite baseball team, the Detroit Tigers. At Crystal, Sooch worked as both a design engineer and as a vice president of engineering and watched the company’s Austin site grow from twenty employees to a thousand.

Fast-forward eleven years, to the summer of 1996. Sooch, Welland, and another MIT grad, Jeff Scott, whom Sooch had met at AT&T, left the office of what was now Cirrus Logic’s Crystal Semiconductor for what was to be a defining lunch. All three were unhappy with the company’s growth and missed the excitement of what was once a smaller, close-knit organization. “It wasn’t as much fun as it was in the old days,” Sooch recalls.

So they drove to Homestead Bar B-Q on Burleson Road and talked about starting their own company. “We had gone back and forth for a while about whether we had the guts to [leave],” says Welland, who, at 45, is the eldest of the trio. “We continued to Jalisco’s, and lunch became beers.” Hours later, they still weren’t able to reach a consensus. “We decided to flip a coin, and it came up heads to do it,” says Welland. The three of them turned in their notices and began having meetings around various kitchen tables about what, exactly, their new company would make. “We knew that it was going to be chips, but what kind of chip, what application, we didn’t know,” says Sooch. “I remember we’d have these planning sessions, which basically boiled down to drinking a lot of wine.” And so, with a downtown office given to them by their lawyers, they did what any other design engineers would do—they started ripping things apart. They dissected cell phones, fax machines, computers, and answering machines, searching for components that could be replaced with computer chips or integrated circuits.

After four months, they found what they were looking for on a modem card—”A big, bulky transformer and messy electronic relays, opto-isolators, and a whole bunch of crap,” says Scott. This collection of gadgetry had a couple of key functions. It provided an isolation barrier, which protected the phone line from electrical power surges, and it translated the computer’s digital language into analog so that it could travel through the phone lines and vice versa. They did an Internet search and discovered that this part of the modem was called the DAA, or direct access arrangement, which was regulated by the Federal Communications Commission and something that, by law, every modem had to have. None of them had ever heard of a DAA before, and they were excited at the prospect of trying to duplicate the process on an integrated circuit, which would make modems smaller and cheaper. And they figured that since no one had made a more elegant DAA, it must be difficult to design. “That’s the kind of thing that we were looking for,” says Sooch. “Things that were really hard, that other people had given up on.” At the same time, Welland had discovered another opportunity hidden in a cell phone—a clunky radio frequency synthesizer that assigns a frequency, or a channel, to each incoming call. Like the ugly DAA on the modem card, he figured that this too could be replaced with a smaller, less expensive chip.

In March 1997, with a $5 million investment from Austin Ventures and two ideas, they were in business. They lined up orders with modem manufacturer PC Tel and 3Com Corporation’s U.S. Robotics unit, the largest maker of computer modems. By the end of 1998, the young company had more than $5 million in revenues. Getting a second round of financing for $7.5 million was easy. By the time the company went public, it had $46.9 million in revenues and was already making profits. Last year the company’s third-quarter revenues alone were $29.4 million, with profits of $6.5 million. The company has since improved the DAA chip, making it programmable to meet international specifications; rolled out the RF (radio frequency) chip to cell phone manufacturers in Europe; and introduced three more products—a chip that allows satellite pay-per-view boxes to send a bill directly to the phone company, a chip that generates ringing at the line’s end at the phone company, and a chip that recovers computer-generated data transferred on the Internet. With nearly a dozen more chips to be introduced this year, Silicon Labs will be that much closer to its rather breathtaking goal—to dominate wireless, wireline, and optical-networking communications with mixed-signal chips and realize $1 billion in revenues in six years’ time. At a growth rate of 50 percent annually, Sooch thinks he can do it. “I think we expected this,” he says nonchalantly. In a rare moment of self-disclosure he adds, “But do I look back and pinch myself every now and then? Yes.”