Some rumors never seem to want to die, apparently including the belief that Apple (AAPL 0.32%) is readying MacBook products with Apple-designed, ARM (ARMH +0.00%) based silicon. The notion of such a device was first offered up in 2010 by Charlie Demerjian, founder of the popular technology news/analysis website SemiAccurate. Nonetheless, as the years have rolled on, Intel (INTC +2.59%) has continued to advance the performance and power of its laptop processors and Apple has adopted them for its Mac products. But could that beautiful partnership be coming to an end?
Apple could do it, but...
Apple is the world's richest and (arguably) most powerful technology company. There is no question that if Apple wanted to, it could go through the hassle of porting all of its first-party Mac OS X applications to ARM, and it could even hand over piles of cash to the third-party software development houses to get all of those applications ported. There is nothing fundamentally stopping Apple from doing so other than the fact that it would take a lot of work.
The question, however, is simply whether Apple today could design a better chip than Intel. While Apple has shown remarkable prowess with its mobile chip designs, offering easily the most powerful and interesting mobile CPU cores, there is a world of difference at both the CPU core level and at the platform level between a processor intended for a smartphone or tablet and a chip aimed at laptops and all-in-one desktops.
...can it do it better than Intel?
In particular, Apple would first need to significantly boost the per-core performance of its mobile processors, and it would also need to step graphics up to another level. Apple can do it, but given that the company (along with the rest of the semiconductor ecosystem) is at least a generation behind Intel when it comes to transistor technology (which directly impacts performance per watt), it is difficult to see Apple being able to match or exceed what Intel is doing with its PC processors. The chipmaker has a manufacturing lead and has superb designs for that space.
Some folks will loudly claim that Intel's manufacturing lead is nothing but vaporware, but consider that Intel plans to have its first 14-nanometer, second-generation FinFET-based PC products in devices by holiday season of 2014 (they're in production now, per Intel). By that time, very few players will be on TSMC's planar 20-nanometer manufacturing technology (Apple will probably be first), with that technology ramping throughout 2015 from large players such as Qualcomm.
Given Intel's manufacturing lead and its repeated demonstration of design leadership in the PC space (Intel's mobile designs today aren't all that great, which is why the 22-nanometer FinFET advantage went largely wasted there), it would be very difficult for Apple to offer products that provided equivalent, let alone better, performance per watt.
Foolish takeaway
It's not wise to say "never" -- Apple is an incredible company, and if it wanted to move mountains to put its own ARM-based silicon into the MacBook, it could probably do it. However, the Mac represents a fairly small portion of the company's overall sales, so Apple would not a really good reason to design a special system-on-a-chip and require a massive rewrite of all software for a roughly 20-million unit market. (Remember, Apple can't reuse an iPhone chip here as smartphone chips do not have all of the necessary IPs integrated for PC use such as PCI Express and Serial ATA.) Given Intel's excellent execution with its PC chips, it's difficult to find that reason.
