Start-ups may be the U.S.' best hope in the high-tech production competition against Japan.
Some people say entrepreneurship is 'crippling' the United States ininternational markets. Oh yeah?
Are entrepreneurs bad for America? Influential thinkers such as former Commerce Department official Clyde V. Prestowitz Jr. and MIT's Charles H. Ferguson argue that they are -- particularly in such key fields as computer memory chips, in which the Japanese control roughly four-fifths of the world market. With new plants for DRAMs (dynamic random-access memory) costing up to $200 million apiece, the anti-entrepreneurs say small companies are just nuisances that drain serious competitors of the very people needed to challenge the Japanese chip manufacturers in the international marketplace.
"Entrepreneurialism, America's pride, may turn out to be an American burden in the country's race with Asian business cultures. . . . ," wrote Prestowitz in the April edition of our sister publication, Business Month. Since the giants of the semiconductor industry represent our only real hope of succeeding in international competition, Pres-towitz suggested, we would be well advised to come up with public policies that discourage entrepreneurship, and that bolster bigger, more established companies at the expense of the small, growing ones.
The argument may sound logical, but you can bet Pres-towitz never tested it on Stan Salva. A 30-year veteran of AT&T, Salva has fought the Japanese on the front lines of the chip war -- and knows what it means to lose. Back in 1984 he helped build and design Ma Bell's state-of-the-art DRAM facility in Lee's Summit, Mo., just south of Kansas City. Less than three years later AT&T came under severe price pressure from the Japanese. Rather than stand and fight, the company responded by ordering that the Lee's Summit facility be shut down.
Missouri is known as the Show-Me state. AT&T's decision certainly revealed something about the big company's resolve to Salva. "It tore my heart out," recalls the normally laconic Missourian. "It was really wrong. If we were patient we could have made it. But the company just couldn't handle the challenge."
Today, the white-haired Salva is back at his old plant, gearing up for a new round with the Japanese, but he's not doing it for AT&T this time. Like most of the large chip producers, Ma Bell has all but quit the business. Instead, he's director of operations at Alliance Semiconductor Corp., a tiny San Jose, Calif., start-up that began leasing the Lee's Summit plant and equipment last April.
Alliance is the creation of one of those entrepreneurs whom Prestowitz likes to rail against. His name is Dan Reddy, and he's a typical high-technology nomad, with career stops at RCA, Fairchild Semiconductor, and Synertek. After years of knocking around the industry, Reddy decided to strike out on his own, launching Alliance in 1985. He soon discovered, however, that few venture capitalists in adventurous Silicon Valley had the moxie to challenge the Japanese in memories. Try as he might, he couldn't even manage to raise as much as $1 million from private investors. That wasn't enough to build a factory, so Reddy sold his company's innovative memory designs to Asian manufacturers, notably Japan's NMB Semiconductor. The proceeds accounted for virtually all the company's $2 million in revenues last year.
But Dan Reddy, like most entrepreneurs, wasn't hankering to sell his future to the Japanese. From the start, Alliance has sought to turn the much-vaunted Japanese advantage of bigness on its head. Instead of going after the "commodity" segment of the market -- where capital expenditure and cheap manufacturing are key -- the company seeks to exploit such entrepreneurial strengths as quickness to market and technological virtuosity. Its memory chips, developed largely by Reddy's brother C. N., a former Texas Instruments Inc. DRAM designer, are targeted at the high-speed end of the DRAM market. That's the fastest-growing segment of the memory business; according to Fred Jones of Dataquest Inc., it has jumped from 13% in 1988 to 18% of the total semiconductor market this year.
Taken in isolation, Alliance's move might be seen as insignificant in the struggle against Japanese domination in computer chips. But the company's focus on exploding niche markets mirrors a trend that's been sweeping the semiconductor business. While such long-established chip makers as National Semiconductor have lost the technological initiative, a host of smaller companies have helped maintain the U.S. lead in many key markets. Seven-year-old Cypress Semiconductor has been wildly successful producing, among other things, state-of-the-art static RAMs and microprocessors. Linear Technology, another young semiconductor company, is on the cutting edge in analog circuits. In the rapidly emerging gallium-arsenide field -- crucial for the production of supercomputers -- two small southern California companies, Vitesse Semiconductor and Gigabit Logic, have taken the post position.
New companies may represent the United States's best hope in the memory business too, despite the huge cost of entry. According to Dataquest's Jones, the two largest domestic producers for the merchant memory market are Texas Instruments and Micron Technology, a company that entered the field in the early 1980s and now registers sales of more than $450 million. Alliance, working out of the once-discarded AT&T plant, will join Micron this fall in picking away at the Japanese memory monolith. Given the huge demand for these chips, Reddy estimates Alliance's revenues could reach $100 million by the end of 1991.
Granted, Alliance alone can't stop Japanese domination of the memory business. The newly formed seven-company consortium called U.S. Memories Inc. could also help -- if we're patient. The consortium's president, former IBM executive Sandy Kane, admits facing big challenges, such as raising $1 billion in capital. Even in the best of circumstances, Kane says, it will take time to put the deal together and get into production -- perhaps two years or more.
Meanwhile, the airplanes carrying computer chips will continue to land, and the Japanese will continue to consolidate their hold on the DRAM market. For the time being, U.S. Memories -- and most of the giants so beloved by the policy gurus -- won't be able to challenge them. Dan Reddy and the folks in Lee's Summit will.