The $1-billion memory chip venture proposed by seven of the nation's leading technology companies is "a step in the right direction" for U.S. industry but must overcome an array of serious obstacles to succeed, experts said Thursday.
Analysts said the venture, U.S. Memories Inc., faces the prospect of fierce competition, technological challenges, private antitrust suits and conflicting interests on the part of the various concerns backing the project. Moreover, they said, the venture must live down the comparisons to other high-tech consortiums that have produced, at best, lackluster results.
Must Get Antitrust Waiver
U.S. Memories, headed by former IBM executive Sanford L. Kane, hopes to build a state-of-the-art plant for the production of the latest generation so-called dynamic random access memories, or DRAMs. DRAMs are used in nearly all computerized devices, and large Japanese firms now control 90% of the world market. Japan's growing dominance of the chip industry has been the source of broad concern in the electronics industry for several years, and U.S. Memories is the most dramatic response to date.
The backers, led by IBM, are Digital Equipment, Hewlett-Packard, Intel, National Semiconductor, Advanced Micro Devices and LSI Logic. They have agreed to invest in the project provided that additional U.S. electronics firms come forward to provide some of the $500 million in equity. Kane said an additional $500 million of debt financing would then be raised. The company also must gain an antitrust waiver and negotiate a licensing agreement for IBM's DRAM technology before proceeding with plant construction, currently slated for year-end.
Ulric Weil, a computer consultant in Washington, was less than enthusiastic about the whole idea. "There's a definite trend in this country to try and solve these high-tech problems by putting together a consortium and getting an antitrust waiver," Weil said. "But what have they produced so far? Very ordinary results . . . and the whole idea leads toward industrial policy and government control."
Even commentators who are not opposed to industrial policy in principle, such as Robert Reich of Harvard University's John F. Kennedy School of Government, have some doubts about U.S. Memories: "American companies shouldn't be dependent (for DRAMs) on the Japanese. But many members of this consortium are also global companies with chip facilities around the world, and it's not clear how this U.S. consortium fits into their global strategies. The second problem, in a nutshell, is that when we relax antitrust curbs at the same time as we erect protectionist barriers, the result is to increase prices and reduce incentives to innovate."
While he favors the U.S. Memories initiative "on balance," Reich said it should be combined with a removal of the protectionist measures embodied in the 1986 U.S.-Japan semiconductor trade agreement.
Most agree that the government is unlikely to block U.S. Memories on antitrust grounds, given the current political climate. But Professor Thomas M. Jorde, an antitrust specialist at UC Berkeley's law school, noted that Justice Department antitrust clearance provides no protection against private antitrust suits, which can be a big threat because they award triple damages.
Jorde said U.S. Memories, which he strongly supports, was "precisely the kind of venture" that illustrates the need for a change in the antitrust laws to provide some protection against treble-damage suits.
But some potential competitors to U.S. Memories see the antitrust question in a different light. T. J. Rogers, chief executive of Cypress Semiconductor, confessed that he was already worried that the new company would start trying to compete in his company's speciality, the so-called static random access memories, to supplement its income from DRAM sales.
"That would be a natural extension of the business . . . believe me, I'm going to be mighty upset if they get an antitrust exemption to compete with me," said Rogers.
Financial analysts were generally upbeat about the company's prospects, assuming it finds the additional shareholders. "They have every opportunity of getting 5% of the market by 1993, and eventually 10%" of the 4-megabit market, said Dan Klesken of Montgomery Securities, who put the venture's chances of success at 70% to 90%. (The 4-megabit chips, which handle 4 million bits of information, are expected to emerge as the dominant memory device over the next few years.)
Times staff writer Paul Richter in New York also contributed to this story.