Some of the biggest U.S. chip makers have been slow to adapt and have found themselves losing markets. That's why they've been screaming so loud for tariff barriers in recent years. But while the biggies struggle, the smaller companies are doing very well indeed, not just by being agile but by leveraging their strengths into a substantial economic presence. In the emerging relationship-based branch of the business, bigger and smaller enterprises support one another, as they do in Japan, rather than trying to outmuscle or outflank one another.
Because of the Japanese, and because of its rapidly changing technology, the semiconductor industry is three steps more competitive than most businesses. But it may not take long for companies in other industries to catch on to the benefits of specialization and relationship building. It's easier to do something very, very well when it's the only thing you do. More and more, today's marketplace requires that you do something very, very well.
In semiconductors, at any rate, the new strategy has proven itself a powerful competitive tool, both for individual companies and for the United States as it confronts the Japanese. Rodgers, who has become a kind of spokesman for the upstarts in the U.S. semiconductor industry, is only too aware of both sides; asked if he expects to see more relationship building in the future, he smiles broadly.
"As an American, I hope so. As the president of Cypress Semiconductor, speaking selfishly, I hope not. Think about it! Intel and Advanced Micro Devices, my competitors, have spent so much money on lawyers. They could each own half of my fab. They could each own half of the four start-ups [sponsored by Cypress], which account for half of Cypress's growth this year. But they don't. All they've got is lawyers' bills marked paid.
"So I certainly hope that Intel and Advanced Micro Devices never learn to cooperate with each other. It's the most potent competitive weapon I've got."
FACTORY OF THE FUTURE
What's wrong with this picture? You are looking at the main manufacturing floor of a $51-million semiconductor equipment company. You count the production workers on your fingers, and you have three fingers left over. Yet this isn't some automated Japanese factory with robots laboring in the dark. There are no robots here, only the seven gowned and booted human beings.
Robert Graham, president of six-year-old Novellus Systems Inc., chuckles at a visitor's mystification. "How did we produce $51 million in shipments last year with seven people? Simple. We buy completed, tested modules from our vendors. We don't allow them in here until a week before they'll be used. In a week and a half, our seven people can put together an $800,000 or $900,000 machine."
Like the new chip companies, semiconductor equipment manufacturers such as Novellus are setting up long-term relationships with suppliers, relying on partners for all but a couple of key modules in the production process. But building and maintaining the necessary relationships is a demanding managerial art.
Vendors must first be qualified and, as Graham puts it, trained. "We look at a shop and see what kind of equipment is in it. We look to be sure they're solvent. We ask if they'll do special things for us." A welding company Novellus wanted to work with did high-quality work but didn't know how to test for vacuum tightness. "We bought them a vacuum tester and showed them how to use it."
When a problem crops up at a vendor's shop, that problem is by definition Novellus's. A laser weld on one part was never even, and the vendor complained that he just couldn't get it right. Novellus engineers paid him a visit -- and ended up buying $10,000 worth of tooling to solve the problem. If a supplier needs financial help to ensure just-in-time delivery, Novellus provides it. "We'll buy the inventory and sell it back to them," says Graham. "They don't pay until it's delivered as finished product."
Novellus's nearly incredible productivity figures -- $365,000 to $380,000 in sales per employee -- reflect the extension of the company's boundaries into its suppliers' shops. But if the productivity figure is partly illusory, the bottom line is not: aftertax earnings of 22%. "And that's in an industry," says Graham, "that's supposed to be sick."
RESURGENCE OF A REGION
Studying Silicon Valley for a master's thesis back in 1980, AnnaLee Saxenian heard what was then the standard line about the region: its growth was over. The big chip companies were building huge plants elsewhere, often overseas. Start-ups could no longer afford to enter the business. Whenever she returned to the valley over the next several years, though, both the region and the semiconductor industry seemed to be bursting at the seams.
"It kept growing and growing," muses Saxenian, now a professor of city and regional planning at the University of California/Berkeley. "Obviously we hadn't heard the whole story." Embarking on another round of research, she discovered the new wave of thriving semiconductor start-ups -- 85 new companies founded between 1979 and 1989, generating some 25,000 jobs and $2 billion in annual sales. She also found that most of these new companies were specialist producers who established close, long-term relationships with suppliers and with customers.
And what made these relationships work? The answer contradicts the common wisdom that with today's technology and high-speed communications, companies can successfully compete no matter where they are located. In many cases, says Saxenian, geographic proximity was key to the relationships. When your supplier or customer is down the street -- within biking range, as one CEO put it -- you can work out the bugs that inevitably crop up in any high-tech product development. Engineers from both companies can get to know one another, boosting the trust and teamwork necessary for collaboration. "There is no good way to [collaborate] if you're more than 50 miles away," one of the CEOs she interviewed told her.
The gloomy prognoses of 10 years ago, Saxenian points out, focused only on the region's drawbacks, such as its high costs. They missed its strengths, such as the range and depth of expertise to be found among companies within that 50-mile radius. Fortunately, the new wave of chip companies didn't.
