The Real Question . . .

AMD is likely to do some slick things in the next week.

They’re going to have some sort of dog-and-pony show in Prague (Why Prague? What connection is there between Prague and AMD? Why not Dresden, and why is this so hush-hush?)

It looks like they’ll talk about 1) 65nm parts and 2) Chartered making some processors for them.

I think everyone, whether media-maker or media-taker, ought to realize that the key question here is not “When will Dresden or Chartered start making 65nm processors?” but rather “How many 65nm processors will Dresden or Chartered make by X date?”

AMD could probably make a nominal number of 65nm processors pretty quickly: FXs, top-end Opteron, thousand dollar chips. That’s nice, but essentially irrelevant to the overall market.

What really matters is the rampup schedule for 65nm parts. Two good milestones would be “When will 25% of total CPU production be 65nm? When will 50%?”

If AMD answers questions like that next week, that will tell us they’re ready to convert, and pretty sure that they can.

If, however, their responses don’t come near answering questions like that, or if they refuse to answer them, well, that tells you the opposite.

We’re just holding AMD to the same standards as Intel. When Conroe and Company was announced, Intel provided rampup figures. Why can’t AMD, and please spare me the “We’re the three little pigs, and we can’t let the big, bad wolf know where we live” routine. If 65nm is such a big secret, why are you having a party about it?

If you think this is being overly harsh towards AMD, just remember this: AMD introduced Opterons in April 2003, and Athlon 64 in September 2003, but actually didn’t make more than a million of them until the end of 2004 (rather less than 25%).

Fool me once, shame on you, fool me twice, shame on me. Sorry, but some of us don’t have the memory of goldfish.

If I had to guess, AMD will talk a lot about how soon the firstborn 65nm chips will arrive, and not talk about when they’ll make lots of them (outside of the “we’ll be 100% converted by X.”)

If I had to guess, AMD will actually get a handful of 65nm out third quarter, a couple handfuls out fourth quarter, and real rampup will occur in 2007.


Be the first to comment

Leave a Reply