Get Real

Prudential issued a sell recommendation on AMD. Some think this is stupid. I don’t.

The only thing I can fault them for is not issuing one for Intel, too.

The reason why the recommendation was issued was because AMD was doing well simply because Intel had shortages, and when the shortages ended, so would AMD doing well.

I think people are forgetting that there’s two elements to judging how a CPU company is doing: how many they sell and how much they’re getting for them.

Sure, AMD has been pretty much able to sell out production up to now, but have we forgotten so soon at what prices they were selling for?

AMD has certainly been able to charge a good deal more overall than they were just a few short months ago, and it’s hard to find anything else but Intel shortages to explain why.

How long can this keep up? Hard to say.

From the production end, probably a while. Intel’s trying to convert over to .13 micron while making humongous .18 Willies. Their roadmaps seem to imply that this situation isn’t going to change dramatically the first half of the year or a bit longer.

AMD is faced with much the same situation, though they don’t have anywhere near the silicon-chewing Willy problem.

In the short-term, demand is going to determine pricing. If demand drops off a lot after Christmas, we may see price moves. If it doesn’t, we won’t.

However, if we look at the second half of the year, the pricing pressure will come from the supply side. Just making most of your processors at .13 micron frees up a lot of production capacity, which you use to make more processors.

This will help Intel more, but between the shrink and the completion of Dresden, AMD expects to be able to jump capacity from 30 to 50 million processors. That’s an extra 20 million CPUs, or over 10% of total world demand, even if Intel doesn’t make a single extra one (which is not too likely).

AMD’s strategy for late 2002 seems to be “we can make Athlons for a lot less than Intel can make PIVs.” They obviously want to make money from Hammer, but serious Hammer production probably won’t come until 2003. Even there, Hammer looks more like a .10 micron chip, just like the PIV was really meant to be a .13 micron chip.

Intel looks like it’s going to fling out a ton of Northwoods running as high as 3GHz (or maybe more) once it gets its .13 micron operation in full swing and AMD releases Thoroughbred.

All this extra capacity could probably get absorbed in 2004 in a world recovering from recession, but it’s going to be around a year earlier than that.

So AMD is likely to go back to firesale pricing (albeit with lower production costs), and Intel will eventually have to follow.

Recession or not, the other question the CPU makers don’t want you to even think about which is looming ever-bigger is, “Why do I need a 3 or 4 GHz machine?” For the average person with a less-than-antiquated machine, there is no good answer

The world CPU market hasn’t actually been doing all that badly the last year. The only sector that really has taken a wallop is the U.S. home market, and I suspect fear of recession isn’t the only reason for it. I think U.S. consumers have been asking that question for a while now.

The Prudential forecast thinks AMD stock will be down heavily a year from now, and they’re probably right. They didn’t say the same for Intel, and they’re probably wrong.

Unfortunately for rational people, when it comes to CPU stocks, the stock market seems to do the exact opposite of what the long-term indicators are. The dumbest thing about the Prudential recommendation might be expecting people to listen to something so far down the road.

Wouldn’t be surprised if AMD stock kept going up for two or three or four or even six months, then all of a sudden, kaboom.

So please don’t treat this as a stock tip. Consider it rather a long-term buying tip.

Email Ed