Ok, it's happened. Apple is switching to x86. (Insert apocolyptic comment here.) But why? My take: it's the laptops, stupid. Laptops are outselling desktops now. They're the way of the future. But building a laptop CPU is tough. You have to maximize performance but keep power requirements very low so that battery life doesn't crash. It's a big engineering challenge, so if you want to be competitive in the laptop CPU market you have to hire smart engineers and give them lots of money to play with.

Now, let's say you're IBM. What do you gain by spending lots of money on laptop chips? Not much. You get some IP, along with whatever dough Apple can give you. The IP might be useful for your other product lines, but the Apple dough is certainly not going to do much for your bottom line considering Apple's low sales volume. It's different for workstations, where the returns are still somewhat low but you don't have to invest as much effort to adapt a POWER chip to the task. So logically, if you're IBM, you're not going to do backflips to develop the hottest yet coolest laptop chip on the planet.

As an aside, what about the 3GHz promise? Couldn't that be the motivation for the switch? Considering the trouble everybody in the industry has had with increasing clock speeds recently, I don't think the failure to deliver a 3GHz part is all that damning. It's definitely not worth the churn that a migration to x86 will cause, IMHO.

Going back to laptops, look at this from Apple's point of view. PowerPC is looking good in workstations, great in gaming consoles, but absolutely dismal in laptops. Meanwhile, Intel's dual-core Yonah chip is shaping up as an awesome mobile CPU. You have two options. You can stick with PowerPC, maybe use the Freescale dual-core G4 to get PowerBooks a bit faster, pray hard that IBM or Freescale becomes competitive, and spin like mad trying to convince people that there's some reason to spend more to get a slower loptop with worse battery life. I'm sure you can imagine how well that would work. Even when the “MHz myth” argument wasn't all spin it wasn't very convincing. This way lies madness and irrelevance.

The only option you really have is to go with a competitive laptop CPU supplier, and that means either Intel or AMD. For the moment, Intel seems to have the better mobile CPU roadmap. Plus, they've always wanted you as a customer, so they might be prepared to make you a sweet deal, particularly in light of the MS defection to PPC for Xbox 360. Having decided that your laptop CPU is going to be x86, the die is cast. Maintaining fully-supported x86 and PPC configurations doesn't make sense in the long term, so the entire lineup goes to x86.

One more thing. I think we'll soon find out that Tiger was rushed out the door to enable this transition to begin. The whole Tiger release felt strange, like it wasn't quite planned. The beta cycle was chugging along like you would expect, then suddenly it's going gold master. Members of the beta program were reported to have been very surprised when they found out a release was coming. Also, and this is admittedly a minor point, the black design of the website and the new packaging somehow felt like a reheated concept from the Panther release. I suspect Jobs connected the dots, realized the x86 transition needed to be started yesterday, and sent down the command to release Tiger right away so that it would have a little time to shine before the x86 news muddied the waters substantially.