DRAM solution will require time, unity
By Mark Ellsberry
The DRAM business can be complicated, especially if competitors unfairly exploit adverse market conditions for their own advantage. Then, too, negative perceptions are being fueled by skeptical observers, who frankly don't fully understand this business. Put these elements together with some added spin from government factions and you wind up with a highly biased story against South Korean DRAM manufacturers.
Over the past six months, industry pundits and self-proclaimed experts have been highly critical of South Korean DRAM makers' large inventories and subsequent price reductions. But when we attempt to reach a practical balance by slowing down production, we get blindsided. Those attacks come from the soothsayers who are constantly waiting in the wings to get industry recognition. They always have a ready, and supposedly highly analytical, view of such actions. However, most often, we hear inane sound bites that hold virtually no meaning and are completely off target.
Let me set the record straight- and in this case, the term "straight" is relative, because the business of DRAM capacity isn't an exact science, as any DRAM supplier can tell you.
To explain it properly, we need to review a little history. Over the past three years, DRAM volume has nearly doubled each year. So it's not as if we can know exactly what the volume is going to be and simply match it. The truth is, DRAM makers have to keep putting some capacity in place to keep up with expected market growth. We run our business based on customers' forecasts, and if they overestimate, we find ourselves with overcapacity.
Recently, due to some very poor planning, we ended up with excess parts in April and May. As a result, the DRAM spot market was flooded with parts, and the price went down too fast. When we cut back on production, which was the logical thing to do, we were accused of pulling a "PR stunt."
On one hand, if we don't put enough capacity in place, we create a shortage, customers get mad, and DRAM makers are accused of playing games to jack up the price. On the other hand, if we increase capacity too much, everybody gets upset, because overcapacity drives prices down and causes the purchasing community more problems than shortages do. And bear in mind that putting capacity in place takes about 18 months, costs well over a billion dollars, and must be added in large increments.
DRAM makers also take it on the chin when they reduce production below capacity. They may lose market share while incurring higher fab costs. So cutting back on production isn't necessarily the answer to the perennial problem in our industry. And there may not be a correct solution given the instability of the various electronic-system end markets DRAM makers serve. DRAM makers can only base their product designs and production estimates on those of their customers. If they're adversely affected by market conditions, so are we.
The only thing left for DRAM makers to do, then, is bite the bullet, shut down the factory for a limited time, and possibly lose some market share. By shutting down fabs for a few days a month, we're doing our best to balance supply and demand and stop prices from falling so rapidly. If we can do that, we'll stabilize price and DRAM capacity. The downside in this scenario is that a competitor can easily take advantage of our cutbacks by increasing production to take our market share, and then we're back to the original problem.
In early 1997, the industry was in the same predicament. We balanced supply and demand. Pricing stabilized. And this worked well for about three to four months. But a U.S. DRAM maker ramped up production too rapidly and destroyed the balance.
If the plan is to work this time, all DRAM makers must play fairly for the overall good of our industry. A rogue player, and senseless criticism from the sidelines, can keep the DRAM business on thin ice.
-Mark Ellsberry is vice president of marketing for the Semiconductor Division of Hyundai Electronics America.
Copyright r 1998 CMP Media Inc.
|
