Yes, emulation isn't ideal. It is just expected to be a stop gap. A stop gap Apple has used several times in the past successfully. So one they are likely to use again. At the risk of sounding like a broken record if you consider Apple's past strategy this all makes sense. Start with a low-end machine where the buyer is less likely to run very demanding software in emulation and they are less performance sensitive. Also bundle with the machine a full sweet of native apps (aka iLife and iWork). Iron out the kinks while developers port over their more demanding software packages.
Yeah, but now that we're in the midst of the age of computer efficiency, it might not fly as well as it has in the past. In a perfect situation, one that assumes the emulator is as efficiently coded as possible, you need a machine that can provide three times the amount of processing power as the target platform. To simplify it to probably a too extreme degree, a program that needs a 1Ghz x86 processor will need roughly a 3Ghz ARM prcessor to run it equally as well. And that's assuming x86 and ARM are equivalent on performance.
That'd mean your ARM Mac will have to run a lot hotter and a lot faster to run that x86 app, which means less battery life overall. If we assume that Apple would move to ARM due to better performance per watt, that would moot their entire reason for doing so. Native apps would run fine, but there won't be as many native apps at first. It'd take at least a couple of years before it'd catch up, and that's assuming it sells well enough to justify the cost of porting a bunch of high end apps to a new architecture.
Also, like I said before, the Mac brand name has certain expectations behind it. This is especially true in these much lauded post-PC days, when people buy laptops with the intentions of doing something specific with it. If they go out and buy an MBA thinking they can run Photoshop or Office with it, only to figure out...hey, they can't yet, it's just gonna piss em off.
Apple will find itself in a situation where a new Mac doesn't sell because people buying Macs these days know what they're getting into. They'll want to wait for the software to catch up before they commit to the new platform, buying older Macs in the meantime, while the people porting the software will wait for sales to justify porting to another architecture.
Things are different these days. Yeah, Apple could do it, and maybe pull it off. It won't be immediately as nice, but it could catch up. Thing is, are the advantages worth the downtime? If you ask me, they're not. It's a lateral upgrade at best. The biggest disadvantage is that there's no real advantage to making the move.
And you are right there is a bit of "synthetic" guesswork to my prediction but not much. ARM is a RISC based processor and there are still high performance RISC based processors out in the wild. The difference is in the past they weren't cost-effective compared to Intel. I think that is changing. I'm not the only one. AMD, Nvidia, others are already looking to ARM as more than a mobile platform.
I'm not saying ARM's bad. It's fine, and would make for a great platform...were it built from scratch. This is something Apple can't do with the Mac, even if they were to introduce an MBA-Lite alongside the normal MBA. With the competition so fierce these days, any setback could cost you your product line. And with the PC scene being smaller and more focused on the middle-high end, the people it's designed to sell to just won't buy it.
Also, From what I understand, Intels are called RISC-likes these days. They use CISC instruction sets, but are built more like RISC processors. I'll have to wait for someone more knowledgeable about CPUs to pop in and explain what that means, cuz I don't quite get it myself. Only that one type of processor doesn't necessarily have as many advantages over the other like they used to.