Milo, Re: "Tell us all the problems as you see it that AMD is doing so poorly at?"
I wasn't the one you directed your question at, but since you ask, let me give you MHO.
Is it Hypertranport?
AMD seems to have done exceptionally well in terms of support. As to technical things such as bandwidth utilization, power, and cost, we'll find those out after products start showing up with it.
Is it CPU designs?
I give AMD full credit for the Athlon design, but that launched in 1999. In 2000, kudos for getting cache on die. In 2001, it's a bit disappointing. Palomino for the desktop is very late, and Palomino for servers is behind in clock speed relative to Thunderbird. Either they could use a new stepping, or they've completely over-estimated how far the K7 core would take them.
Is it process technology?
AMD has a great process technology roadmap, but they have to execute to it, first. They've also pushed .18u farther than anyone else has, so I guess they get points for that, too.
Is it increasing marketshare that's AMD problem?
They've increased market share, but only at the expense of margins. This may pay off in the long term if they can get better market perception, but right now, it's hurting the company more than it's helping.
Is it not enough volume?
Again, volumes are won by merit of extremely low pricing. AMD's only chance of selling out their inventories is by undercutting Intel at least 25% for equivalent speed grades. The fact that they cannot even price according to performance shows their lack of marketing. But to give them credit, it's not easy to go up against the long time perception that megahertz means performance, so I don't blame them for that.
Overall, I'd say AMD's biggest failure to execute is that they have decided to battle Intel on the basis of volumes alone. They aren't trying as hard to win by virtue of their design, or perhaps such a task is not within their ability. Their tactic also includes collapsing the markets where Intel enjoys high margins, such as the mobile and server markets, in hopes that they can wear Intel out first. So far, it doesn't seem to be working. They are even projecting a loss, rather than stabilizing their pricing and losing market share. If this ends up working for AMD in the long term, then it will be worth it. Right now, few investors agree that it is the right thing.
wanna_bmw |
