Issue



Koreas DRAM Inc. meets reality


05/01/1998







Korea`s DRAM Inc. meets reality

George Burns, Strategic Marketing Associates, Santa Cruz, California

At the beginning of July 1997, the South Korean won stood at the then low of 888 to the US dollar. By mid-December the won stood at 1710 won to the dollar: a spectacular fall of 52 percent in less than six months. The repercussions of this collapse are enormous. Not only has the world`s 11th largest economy been brought to its knees, but the world`s lender of last resort, the IMF, is undertaking its biggest rescue ever with a $57 billion effort. Korea`s high-flying semiconductor companies - Hyundai, LG Semicon, and Samsung - will be suddenly grounded. In 1998, their semiconductor capital spending budgets will be slashed and billions of dollars of planned new plant and equipment put on hold, perhaps permanently.

The South Korean economy was a modified command economy. From the 1960s through the 1980s, the government decided what sectors of the economy could grow and directed the bank to provide low interest loans to the chaebol for expansion in those areas. The semiconductor industry was one of these areas, and Hyundai, LG Semicon, and Samsung were the beneficiaries. But now they are the casualties. As a condition of the $57 billion IMF bailout package, Korea must slash the growth rate of its GDP, reform its financial sector, and its chaebol must begin to repay their huge debt. The days of debt-financed fabs are now gone. Korean semiconductor companies are cutting their new fab investments drastically, and have announced delays and push-backs of several of their major new fab projects. Not counting the value of Samsung`s and Hyundai`s follow-on fabs at Austin, Eugene, and Dunfermline, Scotland, the total value of the delayed projects is $6.8 billion (Table 1).

Click here to enlarge image

Effects on equipment spending

From 1990 to 1996, the top three Korean chip companies spent over $21 billion on fabs and fab equipment. During this same period, the Korean portion of total capital spending rose from 9 percent to 22 percent. However, in 1997, responding to falling DRAM prices and vanishing profits, Korean company spending levels fell by 14 percent. The news for 1998 is even more bleak: financial reform and credit tightening will force Korean spending even lower. We expect at least an additional 24 percent decline from 1997`s level to $4.9 billion, the lowest level since 1994.

This downturn in Korean spending may be partially offset by increases elsewhere, particularly in Taiwan and Southeast Asia. In 1997, spending in the rest of Asia was much more resilient than in Korea. In 1998, spending in China, Singapore, Malaysia, and Taiwan will increase by as much as $2.2 billion over 1997`s level. Such an increase will more than offset the decline in Korean spending.

In sum, the implications for chip equipment companies are serious, but they could very well be offset by increases in spending elsewhere in Asia.

DRAM capacity implications

In terms of DRAM capacity, the cutbacks in investment will have the effect of decreasing 64-Mbit capacity, and hence raising DRAM prices. The cancellation of Hyundai`s Dunfermline fab represents a decrease of 20 percent of the company`s planned 64-Mbit DRAM capacity. Likewise, Samsung`s delay of its Keheung fab represents a decrease of 28 percent, and LG`s delay of its Chongju fab a decrease of 25 percent. For the three companies combined, the decrease in capacity between now and 1999 will be 25 percent less than they had originally planned (Table 2).

Click here to enlarge image

Since the top three Korean companies account for over 40 percent of the world`s DRAM supplies, such a decrease in planned investment capacity will have a major impact on DRAM supply and hence, prices.

Conclusion

Until the current meltdown of Asian currencies, the Asian model for the semiconductor industry was generally well regarded. Witness the size and success of the Japanese, Korean, Taiwanese, and

Singaporean semiconductor industries. Witness also the growth of the emerging Chinese semiconductor industry.

The Asia model, if broadly understood to mean government support, is not limited to Asia, however. Governments in the US and in Europe have successfully provided tax incentives and sometimes outright grants of cash for semiconductor companies to build within their jurisdictions. In the case of DRAMs, this has obviously led to oversupply. Prices for 16-Mbit and 64-Mbit DRAMs are at historic lows in terms of price/bit. The cutbacks in planned investment by the leading Korean companies and by Dongbu are a corrective to a fundamental imbalance of governments and companies building capacity and market share without market constraint.

In the early 1990s, Japanese DRAM manufacturers cut back on their spending for new capacity because of oversupply. Korean companies, then outsiders, saw the Japanese cutbacks as a strategic opportunity and added capacity. This allowed the Korean companies to become world leaders in the manufacture of DRAMs within a very short time.

This time the cutback by the Koreans is another strategic opportunity. On the one hand, it could be an opportunity for the Japanese and other DRAM manufacturers to step up their investments and gain additional market share at the expense of their Korean rivals. But if this were to happen, the present oversupply of DRAMs would likely continue. On the other hand, if the industry can just maintain its current investment plans and not add back the capacity that the Koreans are currently cutting, then the growth of DRAM demand may very well catch up with supply before the new millennium. n

George Burns is principal of Strategic Marketing Associates, a high-technology market research company. Strategic Marketing Associates, P.O. Box 1217, Santa Cruz, CA 95061; ph 408/464-2669, fax 408/464-2025, e-mail [email protected], www.scfab.com.