Apple announces move to Intel chips

So much for being pretty sure that Apple wasn't going to roll this one out today. According to everybody's press releases, reports, etc. and what I heard sitting in the Keynote (now available via QuickTime), Apple will begin shipping Macintoshes that are based on the Intel CPU architecture by June of 2006.

What the devil does this all mean?

Well, like I said in my previous post about this, it hardly means anything to most users. Here are the major ramifications as I see them:

  • Apple gets a guarantee that CPU speed and capabilities will never be a selling point against them in the future (at least vs. x86 Windows, which is the vast majority of the competition).
  • Ancient Macintosh software (stuff that requires Classic to run) will not work on the Intel hardware.
  • Apple will have a truly competitive source for CPUs, as long as AMD is also making x86-alikes. Today, there are two licensees for the PowerPC architecture, Motorola and IBM, and neither seem to find it compelling to move the CPUs forward. Sadly, in both cases, the PowerPC has been sidelined to a specialized processing architecture for DSP-style embedded applications--for Motorola, that means mainly embedded communications systmes; for IBM, that means things like their deals to build Cell processors for XBox and PlayStation 3. In all of these cases, the producers will make a lot more money selling specialized chips in the tens of millions a year than they could ever hope to make selling Macintoshes. AMD and Intel, on the other hand, have a lot of reason to continue fighting hard to make the best desktop and portable computer CPUs...they will collectively own the marketplace.
  • Some games and other software will come to the Macintosh that we have not seen before--some will come faster, others will come for the first time. There's a lot of optimization that is done in the code of Windows games that is deep in Intel-land and Mac users could benefit from seeing this code be directly compatible. However, this will eventually lead to x86-only Macintosh code.
  • Windows emulation on the Macintosh will get better--substantially. Microsoft has every reason to want every Macintosh to be an x86 Macintosh... they can sell them a copy of Windows! If you had a Macintosh that could run a copy of Virtual PC like the ones available for the Windows platform (they basically run at full speed), wouldn't you spend 50 or 100 bucks on that? Of course you would. And Microsoft would be happy to charge you that, plus the cost of Longhorn or XP in order to get you to be another Windows licensee. They'd be foolish not to start readying that product now.
  • Other Windows emulation technologies will get much better. Imagine if the largest hurdle to running Windows on the Macintosh (the processor emulation) were to suddenly disappear... it will in about 6-12 months. Based on this announcement alone, emulations such as those by Linmore Systems (which makes a no-frills emulator that is cheap) should get to near-real-time status pretty quickly. There'll definitely be some work that will need to be done in boxing the CPU correctly, but it's something that Apple will likely encourage. Add to that the possibility of a WINE-type emulation of Windows could be done without licensing Windows itself, and could be rootless (doesn't require having a virtual screen like VirtualPC does today).
  • Apple will always have the PowerPC CPU architectures as a possibility, unless they have learned nothing from the past. Apple's been building OS X on Intel processors for the last 5 years, hoping that they wouldn't have to move to Intel, but knowing that they would have no choice if they couldn't see substantial movement from either IBM or Motorola. Now that has happened, though, Apple can keep the PowerPC versions of its software alive, and encourage third-party developers to do the same, for the foreseeable future. Unlike Apple's move from 68000 to PowerPC in the 1990s, they correctly moved to much higher-level APIs first in this transition allowing them to keep both architectures alive with a minimum of effort.
  • Apple can keep supporting current (and future) PowerPC users with upgrades to MacOS X and applications. As noted in the previous point, Apple's now got a factory for building x86 and PowerPC versions of OS X and their applications suites. As such, they can continue to innovate on both platforms, especially once the PowerPC platform is no longer in active development. Chances are, much of the optimization that we will see (such as that dependent upon GPUs running on the AGP interface and those that could be gained through basic software techniques) will be directly applicable to existing machines on both platforms.

So, the sky isn't falling (not yet) and we can all look on with interest as Apple and her developers navigate these somewhat tricky waters.