Given the situation of being late with Palomino, AMD only had two choices. ASPs or market share. They chose to maintain market share and live to fight another day. Despite the losses I think it was the best choice.
They good thing is that they didn't lose any significant share...maybe even gained a little per x-box chips...
Now they have to come through with higher ASPs...which is one reason the Hammer delays will hurt them. Seems this business is just a battle of delays...again..
Right now Intel is flooring AMD...
So what kind of volume is the Pally actually available in at this point? I haven't gone shopping in the past couple of weeks, but I didn't see any on the retail shelves the weekend after release. IMHO, that's not a good sign for AMD and could be indicative of continuing supply tightness. I did, however, see a 1.8GHz P4 on the shelf... albeit with the 845 chipset (*bleh!*) (although once mated to DDR, I think the 845 will be a decent chipset if the SiS chipset is any barometer).
For as long as I can remember, the "continuing story" of AMD and Intel has generally been one company capitalizing on the other company's miscues. AMD gained a lot of market share with the K6-2 because Intel didn't think the sub-$1000 PC would catch on so quickly. Intel took it back with the second incarnation of the Celeron. AMD was stuck at 400MHz with the K6 series because of a manufacturing glitch and Intel ran roughshod all over them. AMD releases the K7 to rave reviews, but doesn't have a lot of volume and can't take advantage of the performance gap between K7 and Katmai. Intel has the 820 - Coppermine - Rambus fiasco, but AMD still doesn't have the volume to compete. Eventually Intel get's it's act together with the coppermine, but then AMD releases the t-bird. The t-bird scales to higher frequencies than the coppermine, whereas the coppermine ends up stuck at 1GHz. AMD starts to pick up some market share. The P4 is released, but at relatively low volume and at speed grades that don't bury the competition. Intel begins to release higher speeds which compare more favorably to the K7 and market share begins to stabilize. AMD is stuck at 1.4GHz and has a 10 month delay on faster parts (a 1.5GHz K7 was promised at the first of the year originally). The P4 slowly begins to penetrate the market. Now the 1.53GHz part is out, but I haven't seen volume yet.
Key questions going forward:
1) Does Intel have good volume on Northwood? If so, then AMD will find itself in a world of hurt.
2) How hard has AMD pushed their gate lengths to get 1.53GHz out of the K7? If they've pushed them very far, then AMD won't get much out of a bulk 0.13 micron process and will need SOI just to keep up with Intel's 0.13 micron bulk process.
3) When will the K8 see the light of day? It has already been paper delayed several times, and is still more than a year away from launch. If it's not released until 2003, then Intel will be close to releasing 0.10 micron parts... The code name "Prescott" with an 800+MHz FSB has started to pop up on the radar. If those roadmaps are accurate, I'd expect to see that part sometime in 2003...
I honestly think 2002 is going to be a tough year for AMD. I don't think they're going to be able to compete with the speed and performance of the Northwood with a 0.13 bulk process. The SOI process may improve their position, but they aren't even sampling yet. With the release of the K8 in 2003, AMD may be able to make up some ground depending on where Intel is with 0.10 micron... |
