Chip Makers Cope With Tiny Market : Technology: Gallium arsenide manufacturers narrow their focus because the industry failed to boom as expected.

In the mid-1980s when gallium arsenide emerged as an alternative to silicon in making high-performance semiconductors, chip makers and analysts predicted that the gallium arsenide market would skyrocket.

Gallium arsenide provided more speed and horsepower to the chips, which run everything from computers to satellites. In eastern Ventura County, a handful of firms sprang up to develop gallium arsenide, earning the region the nickname “Gallium Gultch,” a play on Silicon Valley in northern California where the major chip makers reside.

Technology companies and venture capitalists poured more than $300 million into starting the gallium arsenide firms, some of which were headed by former employees of Rockwell International, which pioneered gallium arsenide work with its own Gallium Gultch plant. Seymour Cray, father of the supercomputer, gave gallium arsenide a high-profile boost by developing his most ambitious machine, the Cray-3, using the new chips.

But the estimates of gallium arsenide’s market--one research firm predicted sales would reach $13 billion by 1994--proved absurdly optimistic. Sales this year will total only about $100 million. Among other things, gallium arsenide turned out to be a much more difficult material to handle than silicon in making chips, and today no one believes gallium arsenide will replace silicon in semiconductors.

Gallium arsenide “is more brittle, for example, than silicon so it breaks a lot easier,” said Spencer J. Brown, president of GigaBit Logic Inc. in Newbury Park. GigaBit, along with Vitesse Semiconductor Corp. in Camarillo, are two of the leading gallium arsenide chip makers. “It took several years for the engineers to figure out how to fully utilize the material. It’s just taken practice, time and practice,” Brown said.

Practice makes perfect, but it doesn’t necessarily make for a profit. Vitesse, founded in 1984, didn’t turn its first profitable quarter until this year’s first quarter. GigaBit, started in 1981, had to lay off workers in some of its early years, was briefly profitable in 1987-88, but is again posting losses.

So the companies are refocusing their visions to adapt to gallium arsenide’s minor role in the semiconductor industry.

GigaBit wants to sell its products to specialized markets, where customers can use the greater speed and power of the chips for certain computers, telecommunication products, testing equipment and other technology.

“We’ve always tended to look at gallium arsenide as a turbocharger to the silicon engine,” Brown said. “Gallium arsenide won’t replace silicon per se, but will give it that added boost.”

Vitesse, meanwhile, is aiming at replacing one segment of the silicon chip market, the so-called “ECL market,” with gallium arsenide. ECL is a type of building-block chip used in making memory and logic chips for many high-performance computers. Current annual sales for that market are $900 million, Vitesse estimates.

Why the fuss about gallium arsenide? The material is an advanced compound made from gallium, a soft metallic element, and arsenic. The material enables the tiny electronic circuits placed on a chip to operate faster, yet consume less power, than those using the traditional silicon material that comes mostly from sand. That means computers and telecommunications systems, such as phone switches, can process or transport data faster than with chips made with silicon.

Initially, gallium arsenide chips cost two or three times more than their silicon counterparts, and for those predicting gallium arsenide’s growth, “the flawed assumption was that people would pay any price for higher speed,” said Vitesse President Louis Tomasetta.

It turned out they wouldn’t, especially since engineers struggled to pack as many circuits onto gallium arsenide--the so-called “integration” that helps computers become more powerful without getting bigger--as they could with silicon.

Brown noted that several gallium arsenide start-ups on the East Coast have since abandoned the business, despite having parents with deep pockets. They included Microwave Semiconductor, a unit of Siemens AG of West Germany; Tachonics, a unit of Grumman Corp., and Ford Microelectronics, he said.

But although gallium arsenide’s outlook has been scaled back, the industry is maturing and it appears GigaBit and Vitesse--each has about a fifth of the worldwide market--are here to stay.

Until recently, GigaBit mainly supplied the chips for the Cray-3, which is expected to debut next year. But after selling Cray some excess production equipment in exchange for stock in March, GigaBit and Cray have nearly parted ways and now GigaBit plans to begin mass production of a line of chips for about 300 other customers.

Indeed, its recent promotion of Brown, whose background is in finance, to president shows how GigaBit for the moment is placing more emphasis on successfully launching production plans than in design and engineering advances.

However, the chip makers still must fight some customers’ wariness about using gallium arsenide. To promote market acceptance of the chips, GigaBit has priced some chips below competing silicon-based models. Vitesse, meanwhile, last year agreed to share some of its expertise with Japanese computer powerhouse Fujitsu, in part to show customers that a major chip maker had given its stamp of approval to the technology.

“The technology is now viewed as a mainstream alternative,” Vitesse’s Tomasetta said.

Because GigaBit, Vitesse and other gallium arsenide companies are privately held, figures on the industry’s size are not routinely publicized. Vitesse estimates its sales this year will reach between $18 million and $20 million, and it sees worldwide industry sales eclipsing $100 million this year.

Brown declined to reveal GigaBit’s financial results, but said one analyst’s estimate of $17.5 million in sales last year was “a bit aggressive.” Nonetheless, he contends that GigaBit is the industry’s biggest player. But Gene Miles, an analyst at Dataquest, a technology research firm, estimates that a third company, TriQuint, a unit of Tektronix Inc. in Beaverton, Ore., is the top player with annual sales of about $25 million.

In any case, gallium arsenide remains a spec in the technology universe. Its sales represent less than 1% of the $25-billion-plus worldwide semiconductor market, and probably won’t gain more than a 5% share in the next few years, Brown said.

Gallium arsenide was supposed to get an added public relations boost with the Cray-3 supercomputer, but the machine’s delayed arrival has forced GigaBit and Vitesse to sell the material on its own.

“It would have made a big difference if the Cray-3 had hit its time schedule,” Tomasetta said. “It still makes a difference, but it’s not going to be the thing that makes or breaks the technology.”

The make-or-break factor in gallium arsenide’s future will be the chip makers’ ability to beat silicon-based chips on both price and performance and establish beachheads in certain markets. But they face the sheer inertia of the 200 companies in the Silicon Valley.

“Silicon is like doorknobs and gallium arsenide is like combination locks,” said Dataquest’s Miles. With gallium arsenide, “you don’t have to carry a key, but you have to remember the combination. It’s the ‘If it already works, don’t fix it’ routine. Silicon is going to be with us for a long time.”