Torching the Obstacles

Yesterday, Macotakara claimed that Apple is currently testing an A5-powered MacBook Air in their labs.

They’re probably right. But in order to release an ARM-based MacBook Air, Apple would have to overcome three obstacles:

  1. ARM-based chips are slower than Intel chips.
  2. Mac OS X doesn’t run on ARM-based chips.
  3. Third party applications don’t work on the ARM architecture.

Although they may seem like deal breakers, they’re really not.

The Intel chip in the low-end 11 inch MacBook Air scores a 2026 in Geekbench and has a TDP of 10 watts. The A5 in the iPad 2 scores a 721 and has a TDP of .5 watts. Apple could put 10 A5 chips in the MacBook Air, use Grand Central Dispatch to run them in parallel, and get more than 3.5 times the performance while using half of the power. Or, they could stick 4 A5 chips in there, get 40% better performance, and draw 80% less power.

When Steve Jobs announced the switch to Intel back in 2005, he said that “Mac OS X is cross-platform by design.” We already know OS X runs on ARM chips — that’s what iOS is. You don’t think Apple has Mac OS X running on ARM chips in their labs? I bet they have for years.

What if Apple released an update to XCode that let developers click one button to make their apps work on an ARM-based Mac? It’s not like there’s a technical barrier stopping them from doing it. If Mac OS X is architecture-agnostic, the APIs developers use to write their apps must be, too.

It seems to me like people are asking the wrong question. The question isn’t “Will Apple release an ARM-based MacBook Air?” because they are going to. The real question is ”When will Apple release an ARM-based MacBook Air?“