The staple diet of Korea

The staple diet of KoreaKorea is reliant on memories for nearly 20 per cent of its entire exports at one time – when memories are down the country suffers. David Manners examines the Korean situation and their opinion of Siemens. Why does Siemens want to make memories?” asks a top Korean chip executive, “instead of concentrating on what it’s good at.” Although Siemens’ relatively low memory volumes mean that the company is not a significant competitor, its presence in the memory business affects the Koreans because, as the only indigenous European producer, Siemens leads the lobbying of the EU on anti-dumping issues. Memories are important to the Koreans because they were worth nearly 20 per cent of Korea’s entire exports at one time. As such, they are crucial to the country’s prosperity and, when the memory market turned down in 1996 and stayed down, it helped precipitate the country’s economic crisis. For Korea, memory is the staple chip product, representing over 90 per cent of their chip revenues. For Siemens memory remains a sideshow – the company’s strategic thrust is towards leveraging its enviably wide systems engineering knowledge to add value to its silicon. But while Korea may hope for Siemens, and others, to pull out of the memory business, it is unlikely that it’s going to happen any time soon – at least in Siemens’ case. Memories – even when loss-making – are valued by chip companies for three reasons: their regular structures make them good for debugging production lines, the competitiveness of the memory market keeps companies at the boundaries of microelectronics technology, and memories are the easiest way to fill a production line. Full production lines are important in the chip business where 70 per cent of the cost of manufacturing is fixed cost. The downside of the memory business hits when the notoriously volatile price turns down – particularly when it drops below cost. With billion dollar factories needing to be amortised, what do you do? It’s like a recession in the airline industry. Flying a full plane costs the same as flying an empty one – so you may as well sell an empty seat for a price below cost because any money coming in helps the reduce the overall loss incurred by the flight. So the memory makers have adopted the only strategy that seems reasonable in the circumstances – to sell their output for what they can get. One (non-Korean) manufacturer jokes: “We take the orders, ship product to customers and they send us a cheque for whatever they like.” For Korea the situation is particularly difficult. Domestic demand is non-existent because consumers aren’t spending in the face of 6.5 per cent unemployment and rising, high interest rates, job-cuts, salary cuts and employment insecurity. The fear is of social collapse and riots. Central to Korea’s recovery is the companies’ ability to export memory. To that end they have to use whatever investment funds they command to best advantage, and their direction is to accelerate the R&D at the expense of adding new capacity. “The market is getting softer and softer. For three years – ‘96, ‘97, ‘98 – everyone’s been losing money,” says S.S. Lee, general manager of the marketing division at LG Semicon. “LG’s strategy is to be technology leader in memory so we are increasing our R&D effort. We’re trying to recruit more R&D people.”
By pushing quickly to more advanced processes like 0.22, 0.20 and 0.18 micron, by having a large proportion of their output in advanced architectures like PC100, by bringing forward new density generations such as 128 and 256Mbit DRAMs, and by moving quickly to using the new microprocessor interfaces – DDR, SyncLink and Direct Rambus – the Korean companies hope to maintain the value of their silicon.

Leave a Reply

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