We’ve mentioned a few times that AMD doesn’t seem to have anything in the cupboard beyond variations on Hammer. Even if Barcelona is pretty decent, just how much can you tweak the Hammer core to stay competitive, especially since Intel plans to finish copying every else good about it (while AMD hasn’t been doing likewise)?
In contrast, Intel has had one major architectural change in the last couple years, and is due for two more (well, maybe not as major) by 2010.
Furthermore, Intel has even provided some reasonably detailed details about the one debuting in late 2008, and if they can afford to give away such secrets without risking financial ruin, surely AMD can afford to reveal they have something cooking.
Unless, of course, the secret is there’s nothing to keep secret.
You know, it’s not like things are going swell for Green lately. It recently just won the biggest S&P 500 loser on Wall St. title for the first quarter of 2007, losing almost 36% of its value in the last three months.
And this is the company that’s supposed to drive Intel bankrupt in a year? Yes, I know most think the man mad, but you know, if you just search and replace the name of the company going bankrupt from “Intel” to “AMD” . . . . :).
Time To Buy A Solution?
AMD has been in this type of situation before. Back in 1996, AMD had just released its first fully-developed CPU, the K5. It wasn’t terribly good. So what did they do? They bought a fledgling CPU company called NextGen and told them to adapt their product to AMD’s standards.
The result was the K6 and K6-2. While they lacked in some features (namely floating-point performance), they were good (and cheap) enough to keep AMD afloat and bought time for AMD to buy in some more CPU expertise (DEC Alpha) to design and build the Athlon.
Maybe it’s time to try that approach again.
Oops, no NextGens around. There is one other CPU company around, though, that makes a lot of processors these days, though, place called IBM. Maybe Hector should buy it, too.
(pause for laughter to subside)
OK, that’s certainly not going to happen, but one doesn’t need to buy the company to buy technology or technologists from it.
Is it possible for AMD to somehow adapt the Power6 and/or Cell designs to x86? Well, maybe, maybe not, probably not within any reasonable length of time.
Well, if the answer is “No,” then what about buying/borrowing/leasing some of those IBM CPU designers to start on a new design Real Soon Now? After all, both the AMDers and IBMers are supposed to be working with each other in Fishkill already, this shouldn’t be too hard to arrange.
You may ask, “Do you think the IBM folks have magic dust or something and the AMD folks don’t?” Well, yes they do. Not technologically, but financially.
AMD is running out of money, and unless Barcelona is Holy Feces stuff, they aren’t going to get more except under disadvantageous circumstances, or maybe not at all.
An overt AMD/IBM announcement about extending their current ties to design will get Wall Street’s positive attention, if for no other reason than IBM could easily buy AMD, and many would take such an agreement as a sign that IBM was at least interested in the idea.
That probably would help AMD’s stock price more than a bit, which would oddly make it more likely AMD could continue as a standalone company by being able to raise the money it needs to continue.
Bluff or not, the prospect of such an alliance would scare Intel a hell of a lot more than anything AMD can do on its own.
What does it do for IBM? Let’s assume IBM really doesn’t ever want to buy AMD (if they do, well, no more AMD money problems).
AMD does pay them quite a bit in their current partnership and such an agreement could yield them even more.
Maybe more importantly, if you view Intel as your major competitor (and the people making the IBM CPUs certainly do), such a move would mean that IBM would get paid for spreading FUD in Blue Man Headquarters without committing to anything. A healthier AMD means a weaker Intel, and Big Blue would certainly would like that in its own competition with Little Blue.
Yes, it’s essentially a stunt with little real, actual benefit. But perceptually, it may be just what AMD needs.