With personal computer demand remaining in the dumps, technology investors are looking for a sign, any sign, that a rebound is coming. One place to look for answers is the market for dynamic random access memory, or DRAM, which is the most common type of computer memory. When no one buys PCs, no one needs DRAM. But when PC demand starts picking up so, too, do DRAM demand and prices.
Recent Daily Interviews
Bank of America Securities'
So far, the chances for a recovery in DRAM prices this year seem slim. In fact, according to
, a unit of Stamford, Conn.-based Gartner Inc., worldwide DRAM revenue will fall to $14 billion in 2001 from $31.5 billion in 2000, bad news for companies that make DRAM, like
. In 2002 the market should start to rebound but it won't be until 2003 that demand could begin to outstrip supply at a rate that will boost DRAM prices from their recent lows, according to Gartner.
recently took time to talk to Richard Gordon, a principal analyst for Gartner Dataquest with worldwide responsibility for memories research. Gordon previously worked as a senior product engineer for nonvolatile memory products at chipmaker
Advanced Micro Devices
TSC: So, let's talk about DRAM. The memory market has had a tough year so far. Recently, Gartner Dataquest lowered its expectations for the DRAM market this year. Why is demand so weak and when is the market going to change for the better?
Well, we've just revised our forecast for this year and we think that revenues will decline by 55%, which is the worst-ever DRAM revenue decline. The previous time when it was that kind of order of magnitude was 1985. There's a lot of similarities actually between 2001 and 1985 in terms of reasons for the downturn. Usually the downturn is caused by overinvestment on the supply side so a
capital spending boom filters through to capacity coming online. So demand is still there but there's an excess of supply. This time around, though, we've seen more of an effect from the demand side so that as well as having ample capacity, we've also got a slow demand side and also the build-up of inventory. So that's causing a push back among semiconductor suppliers.
TSC: What about inventory?
The PC sector isn't that bad in terms of inventory. There's not as humongous a level of inventory there when you compare it with the communications sector for example. The problem with the PC sector is really a unit-growth slowdown in terms of systems shipments. So the inventory has kind of fluctuated at two to four weeks at best to, at the moment, eight to 10, which is not very good. But of course, the second quarter is the slowest quarter. We kind of expected that to be weak anyway.
So when we see an upturn we'll see it first in the PC sector, assuming we get some kind of strengthening on the demand side. And that won't happen probably until the fourth quarter of this year. If we miss that fourth quarter and if the market is still sluggish by the end of the year, we'll probably have to wait until Q3 next year before we see a big uptick.
The reason is that the last major spending from an information technology point of view on PCs in the corporate space was 1999 ahead of Y2K, so three years from then would be 2002. So there's a theory that a three-year cycle would be a time when we would see an increase in IT spending.
TSC: Weak demand is obviously affecting DRAM pricing, with DRAM prices a fraction of what they were last fall. At $2 on the spot market, can anyone make money off DRAM?
You're right. The demand is so weak that bit growth, which is how we measure demand, is very, very low. This year, we project it might even be less than 50% which is very, very low historically. Obviously when you have that kind of weakness on the demand side, we're going to see pressure on pricing and we've seen pricing falling pretty well steadily through the first six months of the year.
On the spot market, it's about $1 for the 64 megabyte and about $2 for the 128. Contract pricing is a little bit higher. We're probably seeing a $2.50 to $3 contract price. At that level, companies are probably able to cover variable cost and make some contribution to fixed cost, but they'll certainly be losing money on a fully loaded basis.
TSC: Are there any signs that prices are stabilizing? What needs to happen for prices to stabilize?
We're going to need to see a much stronger demand side ahead of the Q4 seasonal upturn than we normally see. If we get that strengthening in the market, we may see some stability in pricing, but again, when you go into Q1 and Q2 next year, it's generally a softer market anyway just for seasonal reasons. So we don't think we're going to see any real firmness or any indication of shortages in the market until the middle of next year. That's not to say that prices will continue to decline. They'll probably firm up a little bit in the Q3/Q4 time frame this year and maybe start to soften a bit in the first half of next year before we get some stabilization.
And the reason for that is that we've had a switch off in capital spending so at some point in time, assuming the demand side is still there and doesn't disappear completely, we're going to get supply and demand coming more into balance. And when that happens we'll see more stabilization in pricing.