When rival giants International Business Machines Corp. and American Telephone & Telegraph Co. announced they would give advanced technology to Sematech, the message went far beyond the arcane technology of semiconductors. The word being sent out on Sematech, the newly formed consortium of U.S. chipmakers, was much more basic: Who says U.S. companies can`t work together? The powerful combination of state-of-the-art technology from two big and bitter rivals, whose prestige as researchers is undisputed, cannot be ignored.
Consortium leaders acknowledge the reaction has been strong, and say they are hopeful foreign competitors will not miss the point.
Sematech is made up of major U.S. semiconductor manufacturers, aided by government money, who have joined forces to improve the highly complex manufacturing processes involved in producing the tiny wafers of silicon used in electronic products. They feel that it is Japan`s superiority in manufacturing skills, backed by similar cooperation and government support, that has given the foreign chipmakers a competitive edge.
The group recently picked Austin, Tex., for the site of its experimental factory. There employees culled from member companies will strive to develop and test manufacturing techniques, tools and equipment. Each company will then be able to take whatever discoveries the group has made and customize them to suit its product needs.
Although Sematech says it hopes for more than double the 13 members it has, much of its recently gained clout comes from the stature of those companies that have already jumped onboard.
IBM may be a pro forma member of numerous trade groups and international standards organizations. But Sematech members say the computer giant, which makes semiconductors only for its own use, has played a strong role behind the scenes in getting Sematech on track.
”Of course, Charles Sporck (president and chief executive of National Semiconductor Corp.) initiated the whole thing, but IBM has been a real catalyst,” said a board member. ”IBM has been very effective as a mediator when bickering and disagreements come up.”
”We would like to get the message across that we are serious about making this work,” said Sanford Kane, an IBM divisional vice president who is also his firm`s representative to Sematech. ”We are not just giving lip service to Sematech.”
The 4 megabit dynamic memory chip (DRAM) that IBM will be letting Sematech use as a test vehicle is so advanced that it is not being produced yet in volume. AT&T`s 64,000-bit static memory chip (SRAM) and its complex manufacturing process also is on the cutting edge of technology.
”Both companies are putting an awful lot up front,” said William Dugan, AT&T vice president of microelectronics. ”Hundreds of millions of dollars went into developing these proprietary chips.”
”The Japanese have always said that American companies simply cannot get together,” Sporck said in a recent interview in his Santa Clara, Calif., office. ”Sematech is an important experiment that should serve as an example for other U.S. industries.
”Ours is not the only U.S. industry that has been put at a disadvantage by the manufacturing capabilities the Japanese have developed. None of us can afford the level of duplication that currently exists.”
Sporck hinted that the combination of strong personalities from competing companies has not always meant smooth sailing.
”We absolutely expect a lot of squabbles,” he said. ”We`ve got a lot of strong opinions and a lot of different opinions. When the big decisions are behind us, when we have a chief executive, and all the paperwork involved in releasing the federal funds is finished, then we expect to see the degree of efficiency and cooperation grow.”
After months of silence, Sematech has suddenly surfaced in the public eye again. The group, pursued by a dozen states, chose Austin as the site for its experimental factory just weeks ago, though it had originally hoped to have a site by last fall.
But then nothing has been easy. Sematech was born in controversy, relying on what amounted to patriotic blackmail to bull its way into existence.
The group brushed aside objections about government bailouts and antitrust matters by painting a scenario in which U.S. companies, unable to finance costly research and development because of lagging revenues, fall further and further behind the Japanese, leaving the government with only foreign sources to supply the chips that are the heart of modern defense systems.
Sporck and his followers made it clear that losing the battle of the microprocessor has considerably more ramifications than losing the 19-inch color TV wars.
Sematech`s professed goal is to bring U.S. semiconductor manufacturing techniques up to parity with the Japanese by 1990. By 1993, U.S. semiconductor manufacturers will be the best in the world, it boasts.
But it`s going to cost $250 million a year for the next six years to do it, and the industry, battered during the last three years by the combination of a cyclical slump and the Japanese dumping of chips, can`t do it alone.
To beat the Japanese at what was once America`s own game, the group demanded, and got, $100 million from the Department of Defense, a request that will be repeated yearly. The State of Texas and the City of Austin will ante up about $25 million in cash and contributions in kind. The other half of the proposed $250 million yearly budget will come from the members.
”After the last few years of losses, the ante is a real stretch for some of them,” said one member.
Fees are roughly 1 percent of a company`s gross sales in semiconductor products for a year, with a minimum of $1 million and a maximum contribution of no more than 15 percent of the industry`s share of the budget.
But the biggest levy being asked of members is not in dollars and cents. It`s in brain power. Sematech wants the best and the brightest people these firms have, and it wants them full-time.
”Our biggest concern has been, what kind of people are we going to get?” said Kane. ”Other consortiums told us their biggest problem has been getting only the kind of people that member companies felt they could do without: the third stringers. Sematech won`t work if that`s what happens here. ”We hope that companies will develop the philosophy that the best way to transfer the knowledge gained here is to have our best people learn it while contributing, so they can bring it home again.”
Both IBM and AT&T are sending teams of engineers who worked on the chips to transfer the technology.
Sporck indicated that Sematech has its eye on a chief executive officer, and has approached National Semiconductor about releasing him to run the group. Kane hinted that it could be someone from outside the membership. The announcement will be made within two months.
By pulling out the heavy hitters to announce the types of products it will use in its first phase, Sematech may have been trying to squelch yet another controversy about what type of chip it is concentrating on.
Should it produce at high volume or low? Should it work on DRAMs, an area largely abandoned by U.S. manufacturers, or the more customized ASICs, which many feel is the future?
The arguments stem from a misconception about what Sematech is all about, says Sporck.
”The product doesn`t matter,” he said. ”Our object is not to build a particular product, it`s to improve manufacturing techniques. We have to have a product because we have to build something; we have to have a baseline.”
In fact, the chips that come off the experimental lines will be destroyed. Sematech, the leaders repeatedly emphasize, is not a commercial concern or a joint-manufacturing plant.
”We want to address a series of issues on equipment trends and new generations of chemical and photoresistance processes,” Sporck said. ”We want to look at how we run a shop floor, at minimum cycle time, and maximum yields, and packaging and testing, and a hundred other areas.”
A primary goal is developing more flexible manufacturing techniques that would allow U.S. chipmakers to switch from one product to another more quickly and efficiently within the same factory.
But while the Sematech leaders insist that any discoveries of manufacturing techniques can be broadly applied across the board to even the most customized chip-making process, others disagree.
”Generally, because of the high volumes involved, you either design a factory for DRAMs, or you design it to be a more general purpose factory,”
said Jerry Hutcheson, an executive officer at VLSI Research in San Jose, Calif.
”The Japanese have already solved the DRAM manufacturing problem, and they are working on the general purpose factory. The question is: Do we want to play catch up, do we want to leapfrog ahead, or do want to do something else?”
But Hutcheson insists he does not oppose the group.
”It`s a rallying flag for the industry,” he said. ”It`s easy to criticize, but what are the alternatives? My feeling is `let`s get it started.` If they pick the wrong topics to work on, they can always change downstream.”
