The Try Athlon
The Bull Rebuttal

By Rob Landley (TMF Oak)

Paul is using old data. AMD became interesting only with the release of the Athlon processor design in August of 1999. Before that, I was a bear on the company as well. Now they finally have products to sell that people want to buy, and that they are capable of manufacturing efficiently.

At the same time, Intel has hit a brick wall. It can't produce enough of the chips it wants to sell. For a company whose main strength has always been manufacturing prowess to have problems with processor availability...

That's not a good sign.

AMD's strength and Intel's weakness aren't related, but the coincidental combination is EXTREMELY powerful. The processor industry lives and dies based on one metric: the price-to-performance ratio. On both price and performance, Intel is losing to AMD.

I've been an Intel bull for years, writing for the Rule Maker Portfolio, which owns Intel stock, and owning it personally. But I first noticed AMD's resurgence when I wrote two Rule Maker articles (here and here) comparing the 64-bit designs from AMD and Intel. The result was that AMD's x86-64 design actually made more sense than Intel's, and I did my best to explain why. My investigation got me looking closely at Intel's relationship with Rambus (Nasdaq: RMBS), Intel's high-end yield problems, and how AMD's other designs (Athlon, Thunderbird, and Duron) stack up against Intel's Pentium III, P4, and Celeron. The more I looked, the more dry rot I found under the surface. It's a bit like finding termites in your house: The problem can be surprisingly extensive before you even notice it.

Rather than merely expressing opinions about all this, I invite readers to duplicate my research. First of all, to properly understand this industry, it helps to know how microchips are designed, and how they are manufactured. I wrote a series of columns on both of these topics, which you can find here:

How microchips are made: part one, part two, part three.

Processor design from the 386 to iTanium: part one, part two, part three, part four.

Once those columns have given you the terminology (or if you just want to skip to the good stuff), Josh Walrath's article is an EXCELLENT summary of what's been happening with AMD.

If you're only going to read one article, read that one. If you are going to read more, here's a bunch:

Reviews of AMD's Thunderbird and Duron variants of the Athlon design, plus its x86-64 extension of Athlon to 64 bits, are available from dozens of websites, including Firing Squad, Sharky Extreme, Ace's Hardware, Gamer's Depot, Tom's Hardware, Planet Hardware, Anandtech, CPU Review, and Ars Technica.

Ars Technica's Jon "Hannibal" Stokes does his usual excellent job of explaining the Athlon's design in a way normal humans can understand. The article's a year old now and a bit technical, but still well worth reading.

Tom's Hardware covered Intel's 1.13 Gigahertz Pentium debacle in detail, first noticing that it didn't work, then exploring why it didn't work, then defending their position in the face of Intel's denials, and finally covering Intel's recall of the defective chips.

Here, Inquest looks at Intel's P4 strategy, and isn't impressed. (That article's very detailed, but if you're in a hurry, there's a nice summary at the end.) ViaHardware has its own piece about the Pentium 4, subtitled "A titan stumbles." Sharky Extreme tries to take a positive approach to Intel's new Pentium 4, speculating that it's possible to drill holes in an existing case to mount the mammoth heat sink without its weight breaking the motherboard, although you still need to buy a bigger power supply.

I'd say Intel is substituting brute force for new technology here, wouldn't you? Perhaps new manufacturing technology will address this, but AMD will benefit just as much from that. According to The Register, AMD's immediate strategy is simply to cut prices and increase performance.

Tom's Hardware has several link-filled columns, including "Hammer, P4, Rambus and More," discussing upcoming processor designs from AMD and Intel, and an article titled "A Titan Falls -- AMD Plays David to Intel's Goliath," which gives a variety of reasons why AMD is performing better than Intel.

One of Tom's Hardware's pet peeves is Intel's reliance on Rambus to address some of the performance problems its designs have been suffering. This is a topic I've Dueled on in this very space. I also believe Rambus is a bad choice, and my bear article has links to some research on that front.

Finally, I'd like to make it clear that Intel doesn't have to suffer for AMD to succeed, any more than Gateway needs to destroy Compaq in order to grow. Demand for computer components of all types is at an all-time high, so much so that Compaq's CEO is worried about component shortages. Reuters expects demand to increase. AMD's flash memory business also has an enormous backlog of orders, already enough to sell all its production capacity for the next year and a half. AMD's newest flash memory products have drawn praise from Cell phone giant Nokia.

And high-flying start-up Transmeta wants to license technology from AMD for use in its own products.

Overall, I'd say AMD is in good shape and getting better. It's out from under Intel's thumb and from its own burden of debt and unprofitability. I think it'll do just fine.

The Bear Rebuttal »

 This Week's Duel

  • Introduction
  • The Bull Argument
  • The Bear Argument
  • The Bull Rebuttal
  • The Bear Rebuttal
  • Vote Results
  • Flashback: Costco

     Related Links

  • AMD Discussion Board
  • AMD Snapshot