Rule Maker Portfolio
The Tortoise and the Hare

By Rob Landley (TMF Oak)

BALTIMORE, MD (Sept. 30, 1999) -- The art of Web surfing is being able to wander off on a tangent at the drop of a hat, and to provide your own hat if necessary. Remember, grasshopper: right click, "Open Link in New Window." Minimize something when necessary. Bookmark often.

This may explain why I decided to recommend Intel (Nasdaq: INTC) again for this month's $500 investment after helping a friend set up a small network in his house so he could use his new cable modem from any of his computers. In the process, I needed an updated driver for one of the ethernet cards, so I hit the Web.

The device driver I needed to update was written by a man named Donald Becker, and he's got more than just device drivers on his Web page. The guy works for NASA (which is cool) where he did some of the pioneering work on Beowulf computer clusters (which are also cool -- OK, maybe not to normal people, but geeks like me make Tim Allen caveman noises over this sort of thing).

He also has more information about different types of network cards than any sane human being could possibly keep track of, and deep in the bowels of a list of esoteric technical information even I was only mildly curious about, I learned something interesting to an Intel investor. Apparently, Advanced Micro Devices (NYSE: AMD) used to dominate the 10-megabit-per-second ethernet chip market, but their 100-megabit "PCNet-Fast 79c971" chipset shipped over two years behind the competition, by which point no network card makers were interested. Intel, on the other hand, could and did make its own network cards based on its in-house 100-megabit chips.

Let me repeat that in English. AMD had the lead in an arena Intel didn't dominate, and it simply dropped the ball. The transition to faster networks caught them off guard, and they were so late to market in the new niche that the competition hadn't just planted the flag in the new territory but had nicely manicured lawns with a white picket fence. Meanwhile, Intel bashed its way to success in the same market with brute force and deep pockets, making a complete product it could sell to end-users rather than components with "some assembly required" by third parties.

Now you know where the title of this article came from, as well as the moral: An early lead does not guarantee success.

This brings us to AMD's new wonder-chip, the Athlon. The chip itself seems quite impressive, and has received numerous favorable reviews. It is currently faster than anything Intel is offering, leapfrogging the high-end Xeon chips.

However, Intel has demonstrated prototype chips faster than the Athlon. So why isn't it selling them yet? Because it doesn't have the manufacturing details worked out to its satisfaction. There's a big difference between producing a few dozen prototypes and having 95% of the chips on each wafer you burn work perfectly.

AMD has never had very good yields on its microchip manufacturing, and thus when it has a popular chip it usually has trouble meeting demand. Part of the reason Intel has much better yields is that it won't put a chip out into the market if it's not comfortable with its ability to manufacture the thing. Often, Intel will redesign parts of a completely functional chip to make manufacturing easier and increase the yield of usable chips per wafer burned. AMD prefers to get a flashy, high-performance design to market first, and worry about resolving manufacturing issues later.

So no matter how wonderful the Athlon is, AMD's potential market is limited by the number of chips it can manufacture, which is a small fraction of what Intel can do even when AMD does get good yields. Intel has chip fabrication plants coming out of its ears. Meanwhile, AMD is up to its neck in debt and continuing to lose money. Trying to outperform Intel is an interesting strategy, but not a guaranteed threat to Intel's dominance.

I'll end with some nice links from AMD's hometown newspaper about the Athlon so you can practice that right-click stuff we talked about earlier. Feel free to left-click if you decide to head to the message boards instead. (Those of you on a Macintosh, try holding the button down for a second when you click on the link.)

- AMD's Big Bet
- AMD Warns of Big Second Quarter Loss

- Oak