Page;
To Korea 
Page;Chip Wars
By
Charles S. Lee in Seoul and Tokyo
Far East Economic Review
Issue cover-dated September 23, 1999
For nearly two decades, Samsung Electronics' top mission has been to become the world's greatest maker of memory chips, the silicon strips that helped fuel the boom in personal computers. Now Samsung Electronics is world No. 1 in memory, and the long-awaited recovery from a three-year slump in sales is powering the company to new heights: On July 29, it announced a first-half profit of 1.3 trillion won ($1.1 billion). The news pushed its stock up 11% that day to close at 197,000 won for a market capitalization of 29.3 trillion won, making it the most valuable company on the Korea Stock Exchange. By September 13, its stock price was 250,000 won.
But the mood inside Samsung is surprisingly sober. The gold ring it grabbed is losing a bit of shine. Its 20% global market share in DRAM (dynamic random access memory) chips, the most common kind of PC memory chips, looks less exciting because of their chronic price volatility and prospects of slower growth. Hi-tech seers say the biggest growth and profits in the next decade will be in designing and making powerful chips to run a wide range of Internet-linked "information appliances," from Web television sets to cell phones that can transmit voice, data and video images. And so instead of popping open champagne bottles over their DRAM triumph, Samsung Electronics' top executives are firing up their troops for a new challenge: to make the company as big in these new smart chips as it is now in memory.
Staying on a high-growth curve is only part of what's at stake for Samsung Electronics. It comes down to the kind of role Samsung wants to play in the fast-moving semiconductor industry. As long as it remains focused on DRAMs, Samsung's role will be that of a commodity manufacturer, striving to churn out more and cheaper products while the really innovative work is done elsewhere--mainly in Silicon Valley. The emerging demand for a new family of sophisticated non-memory chips offers Samsung an opportunity to become an industry pace-setter--if it can muster the necessary creative skills.
What are these new chips? Only the holy grail of a much-heralded new era of "digital convergence," in which computing and accessing the Internet will no longer require traditional computers. Computers will be embedded in all kinds of home appliances and pocket-sized devices, thanks to system-LSI, or large-scale integration chips, also known as "system on a chip." A single chip will handle several key functions of computing, such as memory and logic, or signal processing and graphics.
"Digitalization of consumer electronics will speed up the spread of semiconductors," says Lim Hyung Kyu, executive vice-president in charge of Samsung's memory division. This means firms in Asia have a chance to compete. Asia already is the world leader in making consumer electronics devices, from televisions to audio equipment. The segue to making the brains for a new generation of smart electronic devices seems a natural one.
This is such a new, emerging market that no one has it cornered, although Japan's chip makers are also trying to jump into it. But far ahead of both Samsung and the Japanese are the Americans. Reeling from the Asian economic crisis, Samsung and its rivals in Japan all readily admit that they are no match for Silicon Valley's warp-speed dynamism and sheer engineering brilliance. The secret to Silicon Valley's success has been the emergence of hundreds of specialized chip design houses, unburdened by the huge expense of building chip-fabrication plants. These "fabless" design companies can pour all their money into hiring talent while outsourcing manufacturing--often to Taiwan.
Samsung's strategy is to invest big now in system-on-a-chip research, while it has a strong cash flow from DRAM sales. With the global semiconductor market turning around faster than expected, Samsung is on track to earn 1.7 trillion won this year from its DRAM division, according to Warburg Dillon Read.
The chip maker also hopes to draw on the superior manufacturing skills gained over 16 years of producing memory chips. Compared with the challenge of breaking into memory-chip making in 1983, says Lim, "conditions are better now. We are bigger and have more capital and human resources than in our early days."
To be sure, Samsung isn't planning to desert memory chips, which still will account for more than 70% of the company's chip output this year. On the contrary, executives say memory chips will remain the top priority for a long time. After all, it will be years before demand for system-LSI non-memory chips matches the $30 billion global memory-chip market. If anything, Samsung can be expected to solidify its DRAM leadership, as a wave of consolidation sweeps the global memory-chip industry. Out of a field of 17 players a few years ago, analysts predict that eventually only four behemoths will remain: Samsung; its compatriot, Hyundai Electronics; a proposed joint venture of NEC and Hitachi in Japan; and Micron Technology in the United States. The shakeout will likely mean a drop in supply and price volatility, which could raise earnings.
But the only long-term alternative to the DRAM roller coaster is product diversification. Samsung first announced moves in that direction in 1996 but the Asian financial crisis forced it to brake. Word that the campaign was back in gear came in June, when the company unveiled plans to invest $1.2 billion over three years to expand system-LSI research and production facilities.
Samsung already has tripled the number of engineers dedicated to system-LSI research, to about 2,500, over the past three years. Many are Samsung's best brains with proven track records in memory-chip design. The company values their talents so much that it won't allow the researchers to fly in large groups on business trips, lest a plane crash undermine the company's future.
Within the broad category of system- LSI chips, Samsung will focus on digital-multimedia applications, says Lee Yoon Woo, president of semiconductor operations. "Because we'll be concentrating our efforts in only a small segment of non-memory chips, we think we'll do quite well." Samsung says it already has made notable progress in producing logic chips for digital televisions and digital video disk (DVD) players, combined logic-and-memory chips for notebook PC graphic controllers and a core communication chip for digital cell phones.
In another diversification move last year, Samsung bought the technology for the ultra-fast Alpha microprocessor chip for servers and workstations, developed by now-defunct Digital Equipment. Microprocessors are the most important kind of non-memory chips.
Samsung's goal: To more than double annual sales of non-memory chips to $2.5 billion by 2001, from $1 billion in 1998. (Those figures hide the real magnitude of Samsung's challenge, because much of the 1998 non-memory sales consisted of relatively simple transistors and capacitors, rather than complex system-LSI chips. Samsung sold its transistor and capacitor production lines last December to Fairchild Semiconductor of the U.S.) Eventually, it hopes non-memory chip sales will account for half of the company's semiconductor output. But that will depend upon its researchers coming up with a steady stream of innovative designs for specialized applications.
The plan clearly contains a fair amount of bravado. Only two major semiconductor firms--both American--have made a successful transition out of memory chips to something more profitable: Intel, which shifted into PC microprocessors, and Texas Instruments, which moved into digital-signal processors, the chips that drive most digital-telecommunication devices.
What's more, Samsung's Japanese competitors are trying to do the same thing and have much stronger engineering and design skills, analysts say. The world's most sophisticated makers of digital appliances, from televisions to cameras to car-navigation systems, are located right in Japan, putting Japanese chip makers in prime position to team up with them.
"The Japanese have a longer history," says Scott Foster, an electronics analyst at Lehman Brothers in Tokyo. "They have a broader product line and deeper technical abilities in consumer electronics. It will take the Koreans years to compete with the Japanese in logic chips."
On the other hand, the Japanese chip makers--Fujitsu, Hitachi, Mitsubishi, NEC and Toshiba--are trying to diversify from a position of weakness. While Samsung is making money in DRAMs, they are losing big--a combined Yen117 billion in 1997 and Yen300 billion in 1998.
"Samsung can potentially catch up to the Americans faster than the Japanese because it can invest the profits from DRAMs into non-memory chips," says Jon Chong Hwa, an electronics analyst at KEB Salomon Smith Barney in Seoul. Indeed, Samsung plans to continue investing 10% of its annual sales into research and development, with 20% of that earmarked for non-memory chips.
But staying on top in DRAM also takes constant investment. Because of the high-volume nature of the DRAM business, the biggest rewards go to the firm with the greatest production capacity, most advanced technology and earliest market delivery. At the moment, that's Samsung. Like many other DRAM makers, Samsung is racing to develop more powerful DRAMs. It also has plunged into more sophisticated kinds of memory such as flash memory, which functions even when power is turned off and is used in cell phones and pocket-sized digital organizers.
Jonathan Dutton, an electronics analyst at Warburg Dillon Read in Seoul, cautions Samsung against spreading its research efforts too thin. But Samsung's semiconductor chief, Lee, sees no other choice but to diversify. "In digital electronics, the chip itself basically will be the product," he says. "If we don't do system LSIs, it means we can't make our own digital products."
Samsung executives like to tell a story about the company's early struggle to grow in DRAM output. Samsung's first U.S.-made production equipment was about to arrive at a brand-new chip factory outside Seoul, when managers suddenly realized that local authorities hadn't paved the side road leading to the factory. Worried that the delicate equipment wouldn't survive a bumpy ride, they laid the asphalt themselves. The moral: The road to success may be rough. But Samsung knows a thing or two about smoothing it over.
To Page;
To Korea Page;