You can see them over here. Basically, modest decreases for dual-cores and socket 754 Semprons, more than modest decreases for middle-to-upper end Athlon 64s, clearance prices for socket A Semprons, and diddly-squat for the lower-end A64s you’re most likely to buy.
This is good news (or at least will be shortly) if you were looking for a San Diego chip with 1Mb cache, but otherwise, not too much effect.
AMD’s Market Share Goes Both Up and Down
Last quarter, AMD’s server sales jumped up from 7.4 to 11.2%. This is certainly good for AMD, though it should be noted that in the past, Athlon MPs did manage up to an 8% server marketshare, and 11.2% is still rather shy of AMD’s share of the desktop market.
The real test will be what AMD does the next two quarters. If the people actually buying servers are even vaguely close to thinking as well of Opterons as the buzz around here, marketshare ought to be something like 20% or better.
On the other hand, though, AMD’s overall share of the x86 market went down the second quarter, from 16.9 to 16.2% (though much of this was due to Intel selling an untypical number of XBox chips).
This is no big deal, just an indication that AMD isn’t doing a Desert Storm on Intel overall.
The Production/Price Ratio
AMD’s real challenge is going to be over the next six-nine months. AMD stopped making socket A dies a few weeks ago (it normally takes about three months from start to finish to make a CPU), so AMD production will be mostly Hammer core-based this quarter, and entirely so the fourth quarter. This means they’ll have to sell millions more Hammer per quarter than they have up to now.
Can they do that and keep their price structure? Or will prices collapse, like they have in the past?
The answer to that will really depend on what the big OEMs do. If they maintain or increase their purchases of AMD products (and this may be where AMD expects the first fruits of its lawsuits, to “encourage” OEMs to do so and Intel to let them do so), AMD probably can. If that doesn’t happen, they probably won’t.
It’s probably also safe to say that AMD is far more likely to do whatever it takes to preserve its prices than it has in the past, even if that means cutting production.
The price cuts just announced are just a match for the imminent Intel price cuts, so they don’t indicate any particular weakness.
How can we tell in the future, though, if supply overtakes demand? Marketshare will be the best (though not perfect) indicator. In its last conference call, AMD basically said it expected a big quarter, indeed, the execs said they had been stockpiling a little bit to meet expected sales.
If they’re right, AMD’s marketshare should get somewhere in the 17% range next quarter, with better ASPs to boot. If they’re seriously wrong, expect marketshare to stay the same or decrease, with lower ASPs.
Another indicator will be the OEM price on A64s. There’s one or two retailers who seem to selling OEM surplus. If you see their prices drop quite a bit, that will be a sign that all is not well and some serious price-cutting is coming.
I’ve said what the warning signs will be of a price collapse, but how likely do I think there will be one? Well, I certainly wouldn’t count on it, I would bet against it happening, and even if it does happen, it’s unlikely to happen before 2006.
For the next few quarters, demand looks fairly high, and at least Intel supply may be a bit constricted. Unless that picture changes dramatically, AMD should be fine the rest of the year.
That means if you’re thinking about an upgrade, the prices you see now are probably going to remain more or less the same for the next 9-12 months or so.