RAHUL SOOD OF VOODOOPC has a new blog entry saying that it is inevitable that Dell and AMD will hook up. I think he is wrong, and I am willing to put your money where my mouth is.
To start things off, read the
blog entry that started this all out, the rest of this won't make much sense without it. His contention is that Dell needs AMD to be a player again, and because of this, they will have to adopt AMD soon.
Male cow excrement. I don't buy that for a second, but I am much more willing to take Rahul's word than I would a clueless Wall Street mouth breather, Rahul is bright and can reason. That said, the numbers don't support AMD's ability to supply Dell. If you look at the following graph from Daryl Ostrander's June 10, 2005 AMD analyst day presentation, P21 (I can't find it to link to anymore), he lays out AMD's capacity in units.
Now, AMD is working flat out, Q4 and Q1 are completely sold out, Q2 isn't, but that is both because of seasonal lack of demand and Fab 36 picking up, but I hear they might actually sell out anyway. It would not surprise anyone if they do.
The above graph shows that AMD could make 50 million or so CPUs a year in 2005, and in 2006, that hard number rises to about 57, with the tip of the arrow at about 70. In 2007, you are at 70/100 respectively. Assuming the realistic number is closer to the hard line, we have a good idea of what AMD can do, about a 20% production rise in 2006, and another 20% in 2007. This is all in units.
Once again, let's pull some more numbers, greatly simplified, out of thin air. AMD is approaching a 25% marketshare in x86 CPUs, and Dell is a bit higher than 30% of the market. Again, not exact numbers, and a better than realistic case for the sake of argument.
One problem is that Dell does not sell AMD now, so that 25% of the total market is not counting 30% of it that is closed to them. This means, of the companies they sell to, AMD has about a 35% marketshare. If Dell picks AMD up, it is reasonable to assume that AMD will sell about the same percentage through Dell as it does to the rest of the world, so would jump to about 35% of the overall market, or a whopping 40% more than they are selling now.
That means Henri can buy more Ferarris, Hector can do whatever he does for fun, and buy one for Dirk too. The reality is that AMD's numbers won't allow that, they can't make that many chips even if they can sell them. Dell would put them well beyond capacity, forcing massive shortages and allocations. This would make Sun and HP, not to mention the white box vendors, very unhappy, and it would still mean Dell would also not have enough. Dell coming on board would royally screw everyone
and Dell, so why would anyone want to do this again?
The short answer is that it would not. Not only would not, can not and will not, at least not for the near future. Adding to the mix is a resurgent Intel, its new CPUs, due out in Q3 or so will really be good, and Dell knows this. If I have seen numbers, it is a pretty good bet that Dell has them also, and they can see what I do. They have to tough it out for four more months, then it is game on again. So, in summary, I don't buy Rahul's argument, it is rational, but the hard numbers tell me it won't happen. AMD will be Dell free for a long time to come, or is it the other way around? Rahul on the other hand thinks I am crazy, so we made a bet. You can read about that, and how you can participate,
here.
The INQuirer
Threaded Mode
