Prospects Fade for U.S. Memories : Technology: Proposed with a bang six months ago, the chip-making consortium has not attracted a single new investor.
Sanford L. Kane was flush with success last June when he publicly unveiled plans for U.S. Memories, a semiconductor manufacturing cooperative that he would lead. After all, the venture--designed to provide a domestic source of key computer memory chips--was backed by a veritable “Who’s Who” of American high-technology companies, and it promised to solve one of the industry’s most vexing problems: Japanese domination of the memory chip market.
Then reality set in.
Today, some six months later, the proposed $1-billion venture has still not attracted a single new investor and is in danger of folding before it gets off the ground. Two of the three companies Kane thought would be shoo-ins to support the venture--Sun Microsystems and Apple Computer--have declined to join. The third, Compaq Computer, is still mulling over the matter.
Many other large computer and semiconductor companies have also decided not to invest in the collaboration, saying the money could be better spent supporting other U.S. sources for memory chips. Some also note that the key impetus for the venture, a desperate shortage of memory chips in 1988, has eased to the point that a surplus looms in the coming year.
Finally, some entrepreneurs complain, the joint venture threatens the life of smaller chip companies that have traditionally been the source of innovation in the U.S. semiconductor industry.
The criticism has become so intense that Kane, who once figured the plan would be a “no brainer” to sell to the U.S. electronics industry, has extended his organizing deadline, lowered the initial fund-raising goals and turned to Wall Street as a substitute for the financing he thought he would get from U.S. electronics firms.
The moment of truth is coming soon, and although Kane, who quit a vice presidency at IBM to lead the venture, professes optimism, others say they doubt that the venture can regain its momentum.
If U.S. Memories crashes and burns before getting a chance to fly, the implications could be ominous for the U.S. chip industry, its supporters contend.
They argue that the giant Japanese electronics companies, which already dominate chip markets, will increase their lead and make new inroads into other strategically important technologies. These inroads could range from the supercomputers so vital to the defense industry to laptop personal computers, a product line that offers U.S. companies one of the best new business opportunities in consumer electronics in years.
The failure would also leave literally scores of U.S. companies on their own to battle their Japanese competitors, most of which are not only many times larger, but have such advantages as a government-sponsored trade development association and far lower interest rates for the huge sums of money needed to finance industrial expansions.
Finally, the supporters say, allowing U.S. Memories to slip away leaves the American electronics industry vulnerable to the ravages of another severe chip shortage like the one that eroded sales and profits of most computer makers last year.
“They have a short memory of what happened to them last year,” complains Kane, president and chief executive of the fledgling venture. “That’s the biggest problem we have: shortsightedness on the part of big business.”
But critics respond that the plan is actually a cleverly designed and flag-waving rescue of faltering big chip makers who have failed to remain competitive for a variety of reasons, including some bad business judgments of their own over the years. And, pointing to the software industry, where scores of small publishing houses far outdistance their larger Japanese competitors, they argue that concentrating a competitive effort in a single large cooperative isn’t the right strategy for what should be a guerrilla-style battle.
“The cry at U.S. Memories is, ‘If you don’t love me, you don’t love America.’ But that’s just not true,” says T. J. Rodgers, a chief critic of the venture and founder and chief executive of Cypress Semiconductor, a mid-sized chip maker in San Jose. “Any company that is competitive will win. That’s that. Government consortia won’t make a difference. Period. I really do believe that.”
Whether U.S. Memories takes off depends on whether Kane can raise about another $130 million over the next several weeks.
Originally, Kane said he wanted $500 million from electronics industry investors by New Year’s Eve, but in recent weeks he has lowered his goal to $300 million and said he would accept $100 million of that total from institutional investors currently being courted on Wall Street by his investment bankers at Goldman, Sachs & Co.
So far seven of the nation’s largest high-technology companies, including International Business Machines, Digital Equipment, Intel and Hewlett-Packard, have pledged a total of $170 million to the venture. Each also put up $50,000 to cover its organizational expenses.
Kane now says the deal will fall into place by mid-January.
“If we weren’t pretty confident that we could get this put together, we would pack up our tent and go home,” says Kane, who helped organize Sematech, the electronics industry’s other cooperative venture, several years ago. “I expect a significant number of other companies to be added. And the middle of January is as good as New Year’s Eve.”
But getting U.S. corporations to invest millions of dollars in cooperative “save the industry” movements is clearly difficult. Sematech, a government-business chip-making consortium designed to spur U.S. manufacturing technology, suffered a tortured birth process, couldn’t easily find a willing executive for its helm and has yet to meet its own initial goals.
“We’re all too much the cowboy. We like to go it alone,” says David Angel, a semiconductor analyst at Dataquest, a San Jose market research firm. “That’s our culture.”
But there’s another key reason behind the reluctance Kane has encountered: the ending of a severe shortage of dynamic random access memory chips, or DRAMs, the key chips that U.S. Memories hopes to make.
Today, just a year after the shortage crisis peaked, there are more than enough DRAMs (pronounced “dee-rams”) to go around, and prices have fallen to about 15% of their 1988 peak on the spot market.
For some computer makers, that’s enough of a reason to pass on joining the cooperative.
“We have contracts with worldwide suppliers,” explains Scott McNealy, the chairman of Sun Microsystems. “And as long as I have a reasonably consistent supply at a price, I don’t care where it comes from.”
However, during the crisis, which lasted throughout 1988, the Japanese DRAM manufacturers that supply as much as 90% of the worldwide market couldn’t turn out the tiny wafers of silicon fast enough. U.S. customers had to wait in long lines, all the while believing that their Japanese competitors were getting first crack at the slim supplies.
The result: lost sales, plunging profits and renewed antagonism toward the Japanese, who, many high-technology executives believe, took over the DRAM market in the early 1980s and drove nearly all their American competitors out of the business by selling their products at below-cost prices.
Those conditions were the impetus for U.S. Memories, whose first product was to be a DRAM capable of storing up to 4 million bits of data on a single slip of silicon.
But now that DRAM supplies are abundant, Kane and other industry leaders complain, memories of the shortage have faded fast.
“U.S. Memories should happen; but whether or not it will is an entirely different question,” says Jack Beedle, who operates a semiconductor market research firm in Scottsdale, Ariz. “In fact, I wonder if U.S. Memories isn’t aptly named.”
Among the companies still evaluating the business plan for U.S. Memories are American Telephone & Telegraph, NCR, Tandem Computer, Dell Computer and AST Research. Answers from at least the first two are said to be expected shortly.
The timing is critical.
Even starting in early 1990, U.S. Memories will be a year or two behind the Japanese in reaching the market with its first product, a 4-megabit chip capable of storing up to 4 million bits of information on a slip of silicon about the size of a thumbnail. And by 1993, when U.S. Memories expects to be in full production with that chip, the industry, led by the Japanese, will have started to shift to the next generation: a 16-megabit chip.
The progress of technology leaves the venture with only a small window of opportunity in which to begin operating. And if it misses that opening, the market for the venture’s first product will already have shifted.
The shift is already under way, and some industry insiders suggest that the timing is already poor for U.S. Memories.
According to Beedle, 1-megabit DRAMs currently sell for about $6 each, compared to about $40 at their peak last year. And 4-megabit DRAMs, available in limited, pre-volume production quantities, are fetching $60 to $80 each. But the 4-megabit market, Beedle noted, is slow and won’t heat up until prices drop, probably to the $25 range.
But Beedle said slumping computer sales have depressed demand for memory chips, and a glut looms on the horizon to arrive perhaps as early as next year, just as U.S. Memories would be ready to start production.
Despite his outlook, Beedle believes that U.S. Memories is a good idea. He notes that the feast-famine cycles that have afflicted the chip market in the past are likely to revisit it in the future, possibly with the next generation 16-megabit chip now in research laboratories at major U.S. and Japanese manufacturers.
“We’ll go through what we went through last year all over again,” Beedle says in an exasperated tone. “It’s sad that this proposal isn’t getting the support it needs.”
Still others believe that the potential value of U.S. Memories lies in its ability to revive the domestic chip-making industry by sparking renewed manufacturing of DRAMs, the chips considered to be the technology drivers of the entire semiconductor industry.
Because they are so critical to every computer, DRAMs represent 20% to 25% of the total chip market, analysts say. Further, and perhaps more important, the volume production of these chips leads to improvements in manufacturing technologies needed throughout the chip-making industry.
Initially, critics of the plan questioned whether the cooperative would violate antitrust laws. Cypress Semiconductor’s Rodgers, who has been the most vocal opponent, promised to sue if the venture ever started producing products that competed with those made by Cypress.
But Kane says he has received legal opinions with assurances that the venture does not violate antitrust laws. In fact, Kane argues, the cooperative actually promotes competition because it introduces new participants into a market where currently only three domestic companies--none of which are participating in U.S. Memories--now operate.
“U.S. Memories may not be the absolutely right answer, but we won’t know if we don’t try,” says Michael Borrus, director of the Berkeley Roundtable on the International Economy at UC Berkeley. “The venture is important because it’s an experiment to see if we can develop new structural approaches or if we’re simply going to give the market for these chips away to outsiders.”
U.S. MEMORIES’ RECRUITMENT STRUGGLE
Seven of the nation’s largest high-technology companies are investors in U.S. Memories. The problem has been finding additional supporters; the proposed chip-making venture has not attracted a single new investor in the last six months.
Who’s In Advanced Micro Devices Digital Equipment Hewlett-Packard Intel IBM LSI Logic National Semiconductor
Who’s Out Apple Computer Cypress Semiconductor Micron Technoloty Motorola Sun Microsystems Texas Instruments Unisys
On The Fence AST Research AT&T; Compaq Computer Dell Computer NCR Storage Technology Tandem Computer Tandy
