Japanese face hard choices as DRAM sector plummets

Japanese face hard choices as DRAM sector plummets
David Manners We’ve recently seen Korean and American semiconductor companies very publicly demonstrate their pain – the former with UK plant delays (LG and Hyundai) and the latter (Intel, Motorola, National and TI) with massive lay-offs. But what about the Japanese? So far, despite the tremendous erosion in their profitability, the Japanese electronics majors have seemed impervious to the pain of their loss-making chip operations. What’s worst for them is that the domestic Japanese market is the only one of the world’s four regional markets to decline in 1996 (by 15 per cent) and in 1997 (by five per cent), and it looks like declining again this year. If it wasn’t for the Japanese market, the world semiconductor market would have returned to growth by now instead of declining from $150bn in 1995 to an expected $134bn this year. Early in the new year we saw a flash of trauma at one of the majors – Mitsubishi Electric, which posted a loss for 1997 – when company president Takashi Kitaoka resigned. Apart from that, and some cutbacks in capital projects, the Japanese companies have sailed on apparently without reaction to the critical conditions. “Semiconductor recessions have been brutal in the past but the companies continued to do OK, but this time the economy is not supporting them,” said Malcolm Penn, chairman of high-tech analyst Future Horizons. According to Japan’s Economic Planning Agency, which tracks the country’s leading economic indicators, the Japanese economy has contracted for seven months in a row. So the Japanese semiconductor companies have to sink or swim by how successfully they can sell in international markets. And international markets for the product they mainly sell – memories – are the worst in living memory. The DRAM market, worth $40bn in 1995, is expected to hit $14.5bn this year. The simple answer for the Japanese is to do what the Americans did in 1985 – pull out of DRAM. In 1984 the Japanese had invested more in DRAM capacity than they were to invest in any year since for more than a decade. Predictably, the investment flooded the market. In 1985 you could buy four of the then leading technology DRAMs – 64Ks – for a dollar. The industry is said to have lost $6bn that year (when the world market was worth only $30bn), with the Americans losing $2bn and the Japanese $4bn. But, while all the Japanese companies stayed in the DRAM business, all the Americans (Intel, Fairchild, National, AMD, Motorola), except for TI and Micron, pulled out of DRAM. Andy Grove, then president of Intel, recalls in his book Only the Paranoid Survive how he asked Intel chairman Gordon Moore what would happen if the Intel board was replaced by a new set of directors. Moore’s reply was that a new board would get Intel out of the DRAM business. To the company that had developed and marketed the first commercial DRAM, it was a traumatic decision. But Grove recalls that managers further down the line had, on their own initiative, already been loading Intel’s fabs with high value microprocessor wafers rather than loss-making DRAM wafers. By moving to a product (the 386 microprocessor) with a higher level of intellectual property, by cutting out second sources, by jacking up prices and by promoting the 386 as a brand name, Grove transformed the disaster of 1985 into the triumph of 1997 when Intel became the fourth most valuable company in the world assessed by market capitalisation. Now it’s Japan’s turn to make the tough decisions. Beset by ruthless and unremitting cheap competition from Korea and, increasingly, Taiwan, the Japanese semiconductor companies know that in the long run they need to get out of DRAM. But the Japanese majors have a problem that the Americans don’t. Whereas the Americans can pull out of a product area, close the fabs, lay off the people and cut costs – as TI did last week when it pulled out of DRAM – in the Japanese culture lay-offs are still largely unacceptable. What makes it worse is the Japanese semiconductor majors have all the infrastructure of companies with annual sales worth from $3bn to $10bn. It’s difficult to relinquish a third of your revenues while keeping most of your people and infrastructure in place. “They need something to replace the DRAM,” said Penn. “NEC and Hitachi have a broader base than the others and can move to higher value-added products. They have already been successful in doing that. Both NEC and Hitachi can move to system-level products because they have in-house systems-level knowledge driving them.” However some experts do not believe the Japanese will waver in their commitment to DRAM. “I don’t think any of them will pull out,” said Richard Gordon, memory analyst at Dataquest. “I think they see the next upturn as their reward for taking the pain of the past two years.”


Leave a Reply

Your email address will not be published. Required fields are marked *

*