All, __________________ TMF Money Advisor
A few days ago I had a conversation with an owl who is quite wise. But, at the time I thought she had fallen off her perch once too many times.
She was hooting about how everyone is getting shorted on P4 supplies, not only with 2g parts (which I had expected), but also with low-end 1.5 & 1.6g parts as well. Only the couple of speeds in the middle were available, and Intel refused to down-bin anything to supply the slower speeds. This was the part that really threw me, how simply stupid could Intel be? I didn't say anything here because even my pro-AMD biased viewpoint thought that it was a stretch from reality.
Well, yesterday's article about Dell showed me who was right, and I was very wrong.
But, it gets better. It seems that there are mountains of P3 die being stored up. I suppose most of this is slow stuff that will be used for Xbox, so maybe Intel can make a little money recycling their old junk.
The root of the problem seems to be that Intel can't make P4 die to save their lives. Between its huge size, their horribly lame .18 process, and their unyielding love for "exact copy", production is at a crawl.
What is hysterical is that over the past few qtrs as P4 volumes were in the toilet, they could brute force enough capacity to get what was needed. Today, however, with P4 ramp taking off (only because they finally cratered prices & are canceling P3), even Intel's brute force capacities isn't able to keep up. They've created their own nightmare. So, customers are left hanging, even the golden child Dell. Yes, GTW has enough for now, but I believe that is for only two reasons: nobody is buying their stuff & Intel is desperate to keep them sole-sourced.
Allow me to make some projections here. A few months ago, Intel started slashing P4 prices expecting Northwood in the .13 process to meet capacity here in Q4. But, horror, NW seems to still be fighting problems, and wisely got pushed out a qtr. I'm not sure if the problem is process related, MBs, sockets, or what, but obviously there is a problem. That leaves Q4 as a freebie, P3 dying, P4 unavailable, AXP everywhere.
My main question is, has Intel lost their golden touch? The .18 process was a disaster from day one, and now on day 600 (?), it still simply doesn't work. Now, I need a crystal ball to tell me, is the .13 process going to fail in a similar way? Yes, we've all heard Intel say that (1) the .13 process is going better than expected, (2) better than .18, and (3) will be the fastest transition in Intel's history. OK, answers: (1) low expectations? (2) Could it really get any worse than it was 2 years ago?, and (3) only because Intel can't get away from the problems of .18 fast enough.
Look at it another way, didn't Intel recently say that by Q3 or so, all P4 production would be switching over to .13??? Why the rush? Does .18 P4 have such lousy yields and bin splits? And, as importantly, will Intel be going from the kettle into the fire???
Now for the sweetest part of all. Intel is throwing money at everything. A $billion at MS for advertising, another huge chunk at Gateway for 'their loyalty', advertising on every TV channel every hour, cashback bonuses, kickbacks, purks. You name it, they have a handful of money to throw that way. The only question is how much longer can they keep dipping into the bucket before it's empty? I don't know where all this money has been coming from lately, but apparently the rate of spending has accelerated exponentially these past couple of months.
Meanwhile, AMD is surviving just fine, our bucket is staying level. I stopped in Best Buy today, they didn't have a single AXP system, sales guy told me he couldn't keep them on the shelves. A guy was on the 'creation station' trying to configure a P4 system (telling me how last time he got screwed because the system had P600 instead of P800 & performance was horrible), I simply told him that DDR was a much better solution and with AXP gave significantly better performance. A sales guy came over & asked if he needed help, and to my surprise (to the customer's question) he said that he's personally benchmarked P4 vs AXP, and at the same frequency AXP has an easy 15% lead (I didn't correct him about model #'s). A second sales guy who was helping me later, whom I queried, had similar glowing comments to make about AXP. Needless to say, the customer got their $2200 top-of-the-line AXP� P4 systems were all over, I think they need to dust them more often!
I think the OEMs & outlets were hesitant in stockpiling AXP configs due to their fear of poor acceptance of the model # strategy, but I'll bet today that they are frantically building as many as they can. Good thing too, I guess that keeps them from worrying about trying to get extra unavailable P4s� ;D
This is like the 1 GHz announcements last spring, Intel makes the announcements to get press coverage about great performance levels, and AMD gets all the high-dollar sales. Can it get any better?
So many many questions. So far, I have one answer: the Athlon family, .18 and .13, don't seem to sweating anything right now�
Got money questions? Your answers are just a phone call away! TMF Money Advisor puts you in touch with an objective Financial Planner whenever you need it.
TMF Money Advisor