I'm not so much joining in the discussion as publicly recording what I think is going to happen in a few years based not really on this prediction, but the way things are going in general, so that I can point to this post in a few years and either say "I told you so" or "look how clueless I was."
I think this prediction is right, at least in general terms, and while to hardcore geeks it may sound like a terrible idea, I doubt it is, and it makes a great deal of sense to Apple. That said, I expect Apple will continue to sell "pro" systems of some sort based on Intel chips for the foreseeable future, to cover the developer/Photoshop-jockey/video-editor market. They're just not going to sell all that many of them.
This is why the ARM transition will not be like the Intel transition (and remember we're not talking about something happening tomorrow):
For one thing, two years is a lot of time at the rate the ARM architecture has been advancing. Predicting anything about how fast the chips will be in 2013 (or how much Intel will have advanced by then) is difficult.
In the quarter the G5 Power Mac first shipped, back in 2003, Apple earned $44M on $1.7B in sales, and shipped 787K Macs. In the quarter the first Intel iMacs shipped, in 2006, Apple earned $410M on $4.36B, and sold 1.1M Macs.
In the most recent quarter, Apple's profit was $6B--more than their gross in 2006, and almost as much as the entire company's gross for all of 2003--on gross income of close to $25B. They sold 3.76M Macs, and more notably 4.69M iPads and well over 20M small-screen iOS devices. They also have something like $65 billion sitting in the bank, which is ridiculous.
Contrast this with Intel, which in the last quarter was doing extremely well, with gross of $12.8B and net of $3.16B. Or, for that matter, IBM, which had revenue of $24B and earnings of $2.9B.
In 2003, Apple was a relatively small-time player that got IBM to design a wicked-fast custom desktop CPU. In 2006 they were a somewhat larger company mostly on account of selling a lot of iPods, and weren't in a strong enough position to get IBM to do what they needed with the PPC architecture to the point it could compete with Intel's upcoming Core architecture. Today their Mac business alone is three times what it was then, it's the only segment of the PC industry actually expanding, and the company is HUGE--twice the size of Intel, in terms of financials. Heck, they could buy a controlling stake in Intel based purely on that company's market cap with cash on hand.
Further, of all those 25M+ iOS devices last quarter, every single one was running an ARM processor. While nearly 4 million Macs is nothing to sneeze at, Apple's bread and butter is iOS and ARM-based systems. They know them, they control the whole package, and they have an in-house CPU team for the architecture. One that, based on performance comparisons with the Xoom, is doing its job quite well. They've also managed to sell these devices at prices so low other companies are having serious trouble matching them, while maintaing very healthy profit margins.
As far as Apple is concerned--and with good reason--iOS on ARM is their future. There's no reason to stop selling Macs, but the market for console-style computers is not likely limited to handhelds and tablets--there's almost certainly a lot of demand in the bigger-laptop-with-a-keyboard space as well as large-screen desktops. With the rate of CPU power increase in ARM chips, within a couple of years they're likely to be powerful enough to comfortably handle desktop tasks, particularly considering that the average user really doesn't have any use for anything more than a basic dual-core system--everything else is for pros and bragging rights.
So, by way of prediction, I'd assume that Apple will continue to beef up its in-house ARM team, and once the desktop-grade chips are in place leverage that to replace what we currently think of as consumer Macs with beefier, larger-screen iOS based devices (or perhaps some iOS/MacOS hybrid thing to better handle indirect input, since pointing at a 27" touchscreen is ridiculous for more than a few minutes).
After all, Apple could--and very will might--dump a few billion dollars of their hoard into advancing the ARM architecture in some way that competitors can't match, and/or building out chip fab capabilities to keep prices low and availability high. Intel's entire R&D budget for 2010 was in the range of $6B, AMD's wasn't much over $1B, and Apple likes to control their own destiny, so it's not out of the question if they can hire good enough people.
I also bet that they will keep some "pro" machines--perhaps even those that'll keep the "Mac" moniker--in the lineup, for people who want more traditional workstation software, since there's still a lucrative market for that. These will presumably use Intel chips, but then who knows--even Microsoft is working on a version of Windows for ARM.
And outside the gamer market or the relatively small number of people who need or want a virtualized Windows environment, I seriously doubt most people will care. After all, it hasn't stopped them from lining up to buy iPads, and I have NEVER heard even the most ardent Windows fanboy rant about Windows with the same fervor as a half-dozen non-technical people I know personally who love their iPad.
Geeks and old-school Macheads like myself will wail and moan, and Apple won't care. If they did, the iPad would have run the MacOS.
In related news, Microsoft is in trouble.
That would be pretty funny though.
