Re: I concede to your view that AMD will get a lot of revenue this q. But I differ in that I expect it to be profitless revenue. Q1 should be a $400m loss at the least. Same with Q2.
I have given AMD a small chance of profitability this quarter. It will really depend on the market. If the market is simply monstrous, then AMD has a shot of selling out and eking by with either a small profit or a near miss. If the market is anything less than stellar (i.e. meets expectations rather than exceeds), then AMD is in for another $100-200M loss.
Re: That is the transition to 45 nm. From all indications, that will have even worse problems for AMD than 65 nm.
The recent post from Krystal is a good indication that 45nm is slipping. Of course, we all knew that "closing the gap with Intel" was a pipe dream for AMD, even without having to get HK/MG working. In fact, I think AMD will be lucky to see any 45nm parts in 2008. These may in fact slip into 2009, given the issues and difficulties of 45nm.
Re: Extrapolating beyond that is not possible. So we have to revisit at that time.
Actually, it's not that difficult, given AMD's roadmap. Their paper beat rate is to be tick-tock like Intel, but in reality, they just don't have the infrastructure for this. Shanghai will slip along with 45nm, and Bulldozer will slip even farther because it's a new architecture that depends on a robust 45nm process. AMD is saying 2009, but I bet it will be 2010 given the aggressive feature set and energy efficiency claims. In the mean time, I wouldn't be surprised to see an 8-core dual die MCM with 2 Shanghai die in 2009. AMD once called this Montreal in some leaked rumors, and I think it's their Plan-B when Bulldozer starts slipping.
Re: So if AMD will really have 45 nm parts next fall, they should be sending out samples right now. Since it doesn't seem like they even have final 65 nm samples, my guess is AMD will not have 45 nm parts in '08.
I agree for all the same reasons. AMD is scrambling with unanticipated difficulties with 65nm. They are not going to solve these, and then magically have a 45nm process working and ready on the wings. It just doesn't work that way. I wouldn't expect 45nm until 65nm is well mature, they understand all the issues and work-arounds, and they are ready to integrate actual new features, rather than just turning the dials on their machines to the maximum setting and pressing the magic button....
Re: Regarding your rhetorical question above, I think that will be when AMD reports another $billion plus loss for '08.
The relevant word here is "plus".
Re: Q1 should be a $400m loss at the least. Same with Q2.
Q2 is the worst quarter, seasonally, and yet Intel will have ramped 45nm by this time, and have multiple HVM mega-fabs to cover the market. I expect maximum market share loss for AMD in this quarter, and losses extending past half a billion dollars. That's assuming there isn't major restructuring by this point.
The problem is that as mind share deteriorates, people start losing faith at an accelerated rate. I bet there are still a lot of people out there who believe AMD is a technology leader, because they still remember AMD leading a couple of years ago. To the masses, news travels very slowly, and this last year of Intel leadership hasn't hit all ears, yet. AMD's marketing machine has been planting FUD all year about how great Native Quad Core is, and many people and businesses are still drinking the Kool-Aid. It will take a few more quarters for them to hear the word and see the signs (just like it took several years before people lost faith in Intel's Netburst designs). Intel is leading in performance and energy efficiency right now, and Penryn extends that lead. With the size of the disappointment with Phenom and Barcelona, it's only a matter of time before everybody recognizes this.
