The other day, we said the following:
“It’s possible CT technology will be restricted to the high-end just like Extreme Editions.”
It seems that at least some Intel execs agree with that. This Reuters article has the head of Intel’s desktop computer chips business saying that 64-bits is not likely to make much sense for home users until 2006. Apparently, he thinks home users ought to wait until MS comes out with Longhorn. Unspoken (or perhaps just unreported) is the commonly expressed opinion by Intel execs on the subject that you don’t get a CPU that can access more than 4Gb of memory until you’re ready to stick more than 4Gb of memory into your machine.
It is also interesting to note that in the Intel announcements for CT technology, they never mention desktops. They talk about “servers and workstations.”
Let the Games Begin
It seems that Intel is putting itself in the position of keeping the use of
AMD64 IA32e to the minimum while causing the maximum amount of damage to AMD’s financial prospects.
AMD’s recent strategy with Hammers has been to try to skim the cream off the top and make as much money from Hammers as possible. This essentially means enticing current Intel users (particularly Intel server users) away with the promise of a “cheaper than Itanium” 64-bit machine.
CT technology takes that argument away from AMD. Rather than being able to say that AMD has the only 64-bit solution, they now have to prove to a very . . . uhhh . . . biased . . . audience that they have a better 64-bit solution than good old Intel and its Xeons.
The same principle applies for the workstation market. AMD has tried to put most of its products on the high end, again, using 64-bit as the lead to extract the maximum amount of dollars. Intel bringing 64-bit to workstations again makes AMD have to prove before an Intel-biased audience that it is the better choice, not the only choice.
In both cases, this dramatically reduces the potential market AMD had for these chips. It’s a lot easier for Dodge to convince Ford owners to switch when they have something they want that Ford doesn’t have than if Ford has the same thing. That’s true even if Dodge is really better.
Complain as much as you like, but inertia and making safe bets are powerful forces in the real world. Unfair? Sure, but so what?
In any event, if the pickings aren’t so good anymore, AMD is going to have to move downscale one way or another. Indeed, AMD slashed prices on 8-way Opterons (prices that were pretty competitive against equivalent Xeons to begin with) by about half on the higher-end models. This could well be a harbinger of things to come.
You can probably expect similiar pressure on FX and upper-end Athlon 64 prices in the months to come.
The timing of the first Prescott IA32e chips is also a bit suspiciously vengeful, since AMD will still be stuck at 130nm when they first come out.
And of course, should AMD downmarket and do well selling Hammers at mainstream prices, Intel is perfectly positioned to move down with them, especially if the circuitry is already there.
I’m sure more than a few people three-six months from now will be interested in “unlocking” x86-64 in mainstream Prescotts, though I suspect Intel will lock that up in some fuse deep within the chip.
It will be interesting to see if Intel can force AMD back into the bargain basement with these minimal moves.