Lattice Semiconductor Corp. Business Information, Profile, and History

History of Lattice Semiconductor Corp.

Lattice Semiconductor Corp. designs and markets programmable logic semiconductor devices and contracts other companies to manufacture those chips. After struggling through the late 1980s, the company grew rapidly beginning in the early 1990s, particularly after it moved into the market for more advanced, high-density semiconductors.

1980s Startup

Management novices Rahul Sud and Raymond Capece started Lattice in the early 1980s, when the market for semiconductors was red hot. Sud, a native of India, had worked as a chip designer at both Inmos and vaunted Intel. Capece had gained experience raising capital through his job with venture capitalist Ben Rosen. Although neither partner had experience managing a company, they believed that Sud's ideas and Capece's ability to raise investment capital were a winning combination. The pair formed Lattice International Inc. in April of 1983 with the help of C. Norman Winningstad, the founder of the successful Floating Point Systems, a maker of computers and peripherals.

Winningstad was integral to Lattice's startup because he and several of his friends in the Portland business community fronted much of the initial investment capital. They invested in the company partly because they believed that Lattice's success would help Portland become a U.S. technology center. Winningstad helped Sud and Capece to raise about $19 million. But that sum paled in comparison to Sud's grandiose business plans. Sud decided that the fledgling Lattice should immediately begin construction of a $100-million, cutting-edge manufacturing facility. Scheduled for completion by 1986, the facility would, according to Sud and Capece, churn out high-tech chips designed by the semiconductor-industry superstars who would comprise Lattice's work force.

Sud's and Capece's vision never materialized. Part of the problem was that semiconductor markets slumped in 1985. But even before the industry tailspinned Lattice was clearly headed in the wrong direction. Sud and Capece did lure some top chip-design talent to their company, but they managed the company poorly. The most glaring flaw was their unwise use of the company's cash. Rather than carefully investing the maximum amount of capital in research and development, they squandered money, leasing an extravagant 140,000-square-foot building, and catering expensive breakfasts for the employees. One worker was even given a Porsche for Christmas. The company's posh, fake-marble lobby was enough to turn one investment banker on his heels.

Moreover, Lattice's production schedule began to slide and the company started losing huge sums of money. Lattice's first product had been a promising high-speed memory chip. But the device was introduced early in 1985, in the midst of the industry slump, when few buyers were willing to risk the switch to a new chip design. Also in the works was a high-performance programmable memory chip. But that product ultimately posed too great a challenge for Lattice's design team and was never introduced in finished form. To help buoy lagging sales, Sud decided to try selling different versions of the memory chip. To that end, he hired a giant 65-member sales and marketing team (for a company that was generating sales of only $1.5 million per quarter).

Sud was perplexed by his company's inability to make money. During a trip to the Far East, he became convinced that lazy employees were the problem. In the Fall of 1985, therefore, he moved Lattice to a six-day work week, similar to that in Japan and Korea. "We were working seven days a week," recalled David Rutledge, product development director, in Forbes, "and then they mandated six days." Lattice lost $7 million in 1986 from sales of the same amount. Despite that deficit, it looked as though the company's fortunes might be changing. Sales of a version of its programmable memory chip were surging and overall revenues were climbing. But that temporary boon was squelched when Monolithic Memories, a Silicon Valley chip maker, filed suit against Lattice claiming patent infringement. Sales of Lattice's promising chip quickly dried up and the company found itself back at square one.

By 1987 Lattice was on the ropes. Desperate, Sud scrambled to raise $10 million in venture capital to keep the enterprise afloat. Just one year earlier he had explained to The Oregonian that Lattice would succeed because there were no semiconductor-industry venture capitalists involved to "force-feed the company with their conventional wisdom." Not surprisingly, Sud was unable to secure financing. Some of Lattice's employees paid for critical supplies out of their own pockets and went without paychecks to keep the venture moving. The company was dealt a nearly lethal blow when Seiko, of Japan, stopped producing Lattice's chips. Lattice's board of directors finally sent Sud packing (Sud later filed for wrongful discharge, but settled out of court after Lattice countersued). They brought in Winningstad to try to turn the company around, but it was too late. In July 1987 Lattice filed Chapter 11 bankruptcy to get protection from its creditors.

Renewal

Lattice posted a net loss of $8.5 million for the year, and many analysts wondered if the company was worth saving. Winningstad believed it was. He realized that Lattice possessed some great talent and had developed some promising technology, but it lacked management expertise. By selling stock to an insurance company, he raised $7.5 million to help pay some bills. He also quickly lowered company overhead by slashing the work force, moving to lower-cost facilities, and eliminating other unnecessary expenses. After getting creditors to restructure the company's debt, Lattice emerged from bankruptcy after only 88 days.

Having kept the company from going under, Winningstad now faced the formidable task of making Lattice into a profitable competitor in the semiconductor industry. Success hinged on the company's ability to parlay its technology into products that it could market. When Winningstad took the helm, Lattice was trying to support five product lines. He decided to dump all except the most promising one; General Array Logic (GAL) devices. Lattice's GAL devices were low-density chips used primarily to link other microprocessors in consumer electronics and computers. Lattice's GAL chips were low-tech in comparison to some of its other products. But insufficient capital would force Lattice to shelve work on more advanced technology, such as its electrically erasable memory chips and digital signal processing devices.

Winningstad's most pivotal move at Lattice came in 1989, when he convinced Cyrus Tsui to become president of the company. Tsui was a native of China. He left Shanghai in the 1960s to attend the University of Southern California, from which he graduated in 1968. He worked a brief stint at semiconductor powerhouse Fairchild Semiconductor before getting Masters degrees in electrical engineering and business at Stanford. He eventually went to work at Advanced Micro Devices (AMD), a chip-technology leader in Silicon Valley in the mid-1970s. He bounced from AMD to Monolithic Memories and back to AMD during the 1980s before he was offered the top job at Lattice. Tsui initially rejected Winningstad's offer, but reconsidered.

A New Direction

By the time Tsui assumed the presidency, Lattice had already achieved an impressive recovery from its 1987 low. Sales had grown from $14 million in 1988 and would hit $21.5 million in 1989. More importantly, the company was generating positive cashflow and would record its first surplus in 1989, when it netted income of $2.2 million. But Lattice was still depending on a relatively limited product line and was in need of a long-term growth strategy. To that end, Lattice went public with stock offerings late in 1989 and in mid-1990 that raised about $40 million. Tsui planned to invest that cash in the research and development of new technology, prompted by Advanced Micro Devices' announced plan to target Lattice's niche in low-density programmable logic devices that incorporated cutting-edge complimentary metal oxide silicon (CMOS) technology.

Tsui focussed Lattice on the market for high-density programmable logic devices. The market for high-density chips was growing in the wake of the introduction and popularity of more complex computing and telecommunications devices. High-density chips were used for data-intensive applications that often required reprogramming of the chip. The greatest barrier to entry to the growing industry niche was technology. But Tsui believed that Lattice possessed the technical acuity to excel in the high-density arena, and that success would mean significantly greater sales and, possibly, much higher profit margins.

Relying on sales of its low-density chips, Lattice managed to increase revenues to $38.9 million in 1990 and then to $64.5 million in 1991 (year ended March 1991). Net income, meanwhile, rose to more than $10 million in 1991, providing a much needed boost to the company's bottom line. Then, in March of 1992, Lattice introduced its first high-density devices. The company unveiled a family of eight high-density programmable logic devices along with software tools that chip designers could use to integrate the semiconductors into their systems. The chips were introduced three months ahead of schedule, and Lattice's stock price had grown more than 70 percent since January in anticipation of the success of the new line.

Lattice's new line of high-density devices was well-received, and helped to establish Lattice as a technological contender in that market segment. Bolstering the success of that new line in 1993 was Lattice's $19 million purchase of QuickLogic Corp., a Santa Clara, California, designer and marketer of field-programmable gate arrays (FPGAs). FPGAs were a rapidly growing segment of the semiconductor industry that complemented Lattice's drive into high-density programmable logic devices. FPGAs were more versatile and generally more powerful than high-density programmable logic devices and were typically used in the most demanding military, aerospace, and industrial applications. The merger worked well because Tsui had worked with the founders of QuickLogic when he was with Monolithic Memories.

Lattice managed to boost sales of both its high-density and low-density devices during the early 1990s. Importantly, it advanced in the high-density market rather quickly. That speedy progress was largely attributable to an important advantage; unlike its competitors' high-density programmable chips, Lattice's semiconductors could be reprogrammed without unplugging them from the systems in which they had been installed. As competitors hustled to copy the innovation, Lattice worked to pioneer new advances. Lattice continued to trail industry-niche leaders Advanced Micro Devices and Xilinx (in terms of sales volume), but it made big gains and was closing in on those opponents by the mid-1990s.

Lattice's sales reached $100 million in 1993 and then climbed to $144 million by the fiscal year ending March, 1995. Net income tracked those gains, rising to $27 million in 1995 and then to $41.8 million in 1996 (from sales of nearly $200 million). By 1996 Lattice was employing about 500 workers, serving more than 400 customers, and supporting offices in Europe and Asia. In 1994, in fact, Tsui had connected with his native country, China, when he opened a Lattice research and development center in Shanghai. Tsui planned to target China for expansion because of his intimate knowledge of the country. Going into 1996, Lattice continued to act primarily as a development and marketing company, but was moving to add manufacturing operations in Taiwan.