Intimate Relations

Global pressures are forcing even the fiercest competitors to do something that all the industrial policy in the world couldn't: cooperate

Let's start with three propositions.

First: the marketplace today is intensely, brutally competitive. No doubt that's a cliché; even so, it's worth remembering how much has changed in recent years. American companies in the past rarely had to worry about imports, let alone compete with foreign-owned businesses operating here in the United States. They could assume that the technologies of business -- production machinery, communications and information systems, and so on -- would change only slowly. Those were the days.

Second: nearly all the management advice you get -- yes, even from this magazine -- merely qualifies you to enter the race. Of course you have to cut costs, boost quality, pay endless attention to customer service. But high-quality goods and services at a reasonable price are no more than a ticket of admission to today's marketplace. To succeed, you need another kind of competitive edge.

Third: year after year, the only reliable edge is continually leapfrogging the competition. Jumping into new niches, or into next-generation products and services. Improving your internal efficiency by orders of magnitude. Learning to serve your customers in different ways from everyone else. Notice that word "continually." Recent history is littered with the carcasses of businesses that leapfrogged once, then forgot to look behind them.

Are you with me? If you accept these propositions, I'll offer you an unusual hypothesis, then back it up with some pretty startling evidence from the semiconductor business.

The hypothesis is this. Companies that are the most successful at leapfrogging -- companies that literally build themselves around constant learning and constant innovation -- don't do it alone. On the contrary: they are embedded in webs of the most intricate business relationships you'll ever want to see. They set up long-term strategic alliances, not with just one partner but with several. They depend on other companies for critically important operations, and for help in developing new products. They share people, equipment, and information -- even with competitors -- in ways that would appall traditional managers. At times, it can be hard to tell where one company ends and another begins.

If there's an industry more competitive than semiconductors, I'm not sure what it is.

Semiconductors, of course, is shorthand for integrated circuits, the silicon chips that make possible desktop computers and a hundred other miracles of modern technology. Chip makers take it for granted that today's products and processes will be obsolete in a few years. They also take for granted a formidable array of potential competitors, from giants such as Intel and Fujitsu to the latest high-glamour, venture capital funded start-ups. The competitive struggle takes periodic tolls. Most U.S. companies, for example, have been forced out of the biggest memory-chip markets. Advanced Micro Devices (AMD), a billion-dollar company, recently had to sell one of its production facilities to Sony.

And yet: despite the intensity of the competition, despite the recurrent wails in the press about the imminent collapse of Silicon Valley, plenty of semiconductor companies are thriving. Consider Chips & Technologies or Cypress Semiconductor, each with projected 1990 revenues of more than $200 million and aftertax earnings in the 15% range. Chips is only five years old, Cypress seven. A dozen other young companies, though not so big, are both well established and remarkably profitable. Weitek, which specializes in chips for math-intensive operations, did $49 million in sales in 1989, with net income close to $7 million. Altera, a maker of programmable logic devices (PLDs), was up near $60 million sales and $11 million net. Weitek grew 40% last year, Altera 55%.

The key to this growth has been innovation. Cypress introduced 56 new chips and chip subsystems in 1989 alone. Altera pioneered a new family of PLDs five years ago; came out with a second-generation product cluster last year; and expects a third generation by 1992. That's fast even by Silicon Valley standards. Japanese customers, among others, are lining up to buy. Altera does 15% of its business in Japan, Weitek about 6%, and both companies are expanding their presence there. Novellus Systems, which makes chemical vapor-deposition equipment for semiconductor wafer fabrication, expects to do 30% of its sales in Japan in 1990 (see "Factory of the Future," page 5).

The people running these businesses, in short, are successfully steering them through one of the most perilous marketplaces in the world. But what's interesting isn't their success, it's their secrets. As experts such as AnnaLee Saxenian of the University of California/Berkeley are discovering, they've created a new model of what a corporation is and how it goes about its business (see "Resurgence of a Region," page 5).

In the past, the leading companies in the semiconductor industry were such giants as Intel and AMD. They looked like most big, integrated manufacturing enterprises: a research-and-development division, a manufacturing division running huge production facilities (fabs, in industry lingo), a big sales force pushing a broad product line. They also had the usual American keep-your-powder-dry attitude toward others in the business. Suppliers were just vendors, to be handled by purchasing agents. Customers were taken care of by salespeople. The competition was, well, the competition.

"Our business is tough, and silicon companies tend to be very adversarial, both with their customers and with each other," says Cypress founder T. J. Rodgers, a 15-year industry veteran.

The traditional industry practice of second sourcing -- that is, licensing out proprietary products so customers can feel confident their supplies won't be interrupted -- typically intensified rather than mitigated the competitiveness. "Even though you're in a [second-source] partnership," says Rodgers, "you do everything you can to prevent the other company from being successful." The licensing agreements between Intel and AMD, he points out, have been marked by lawsuits on both sides.

The upstarts, by contrast, stand both the traditional strategy and the traditional attitude on their heads. Instead of establishing broad product lines, they focus on niche markets. Instead of trying to do as much as possible themselves, they specialize in one part of the production process. Weitek and Altera concentrate on chip design and new-product development. Cypress, though it has a sizable product line of its own, pours resources into its leading-edge manufacturing capability and relies on other companies to provide additional new products through second sourcing. A company called Orbit Semiconductor does nothing but manufacture chips to others' specifications -- and has doubled its revenues in two years, to more than $21 million.