doesn't make any technological sense to do this

The only reason I think it makes sense is that it allows Apple to move away from relying on third parties to design their chips. Right now, design, support and brand really are the main things that let Apple machines stand out. If Apple were to develop their own chip in-house, and let it do something no Intel chip could dream of, they'd have something special. Apple didn't buy PA Semi for nothing.

Pushing state-of-the-art in CPU design costs way more in R&D than Apple has historically spent. The purchased PA Semi group doesn't even really design the ARM chips branded as A4 and A5: those are just licensed core and memory designs packaged together by Apple.

http://www.ifixit.com/Teardown/Apple-A4-Teardown/2204/1

Can't they just buy Intel? Or at least AMD? :)

Sure, but that doesn't mean they haven't been working on something behind the scenes for the past years. I could see Apple moving into a position where they are no longer relying on others to make their chips.

Aside from they haven't spent enough on R+D to actually be doing that behind the scenes... sure.

I'm sure Apple could with that huge pile of cash - but I see no business reason why Apple would pump the money into processors necessary over the long run to keep up with Intel/outside vendors. That's part of the reason PowerPC failed and despite the Apple market share increase in the intervening time they still don't have the scale necessary to make things work.

If there's a post-PC point it's not here yet. It won't be in 2013. Perhaps Apple wishes to alienate the core creative users who stuck by them for decades by letting the products core users are interested atrophy - some are already irritated.

Make, design the SoC, or design the cores? With the exception of very-high-margin top-of-the-line desktop chips and a few specialty products, fabricating the actual chips is a high-volume, low-profit, capital intensive business. NOT the sort of thing Apple wants to do.

Design the chips at the SoC level for handset and tablet level stuff? Definitely, and there are lots of signs they've moved in that direction.

Design the cores? Possibly, although it would take some very new ideas for it to be a winner.

There's no way they can compete with Intel at the high end, barring some truly genius idea that actually works out in execution (and the industry is littered with companies who thought they had one, and where it didn't; the only company that's even managed to break even at that was AMD.) The combined resources of IBM, Motorola and Apple weren't able to do it, or even come close - part of that was marketshare-wise on their instruction set (something that Intel-clones didn't have to worry about) but even speaking technically while briefly at some generations they were able to hold even (and for a few brief times, were faster) for the most part, Intel smoked 'em. Heck, even Intel's own bets against its traditional x86 designs have always lost.

I'm not sure the same holds true for ARM, but there would still be the whole installed base issue.

Most of the potential value add is not in the CPU core.

What I think will happen is Apple will keep Intel processors for Heavy Weight duties such as Photoshop and ARM for less taxing duties. I can't see Photoshop or any other highly demanding application running on an ARM processor. In addition, Apple recently made the jump to 64 bit computing, it will be awhile before ARM processors are 64 bit capable.

if by recent you mean eight years ago in 2003, with the powermac g5.

I forgot about the G5. I was thinking about Intel. But in any case, it's somewhat recent in the grand scheme of things.

PowerPC isn't ARM; they're both RISC instruction sets, but still quite different.

i didn't say powerpc was arm.

what i said was that apple had 64 bit processors eight years ago and for computers, that's a long time.

If Apple is bringing IOS and MacOS together, one could envisage support for two processor architectures under the new IOS/MacOS.