Never.
a) you'd need to recompile so many OSX apps that currently only support Intel Binaries.
b) An A5 or A6 would still be far far away from the performance of a Ivy Bridge CPU. Just web rendering is a big task for an ARM. It cannot handle a Desktop workload yet. Even if it would suffice for some few people it would be such a huge step down, I cannot see them do it.
Software optimization can only do so much, also it is doubtful anybody can truly compete with the compiler whizzes at Intel.
ARM will stay iOS devices. If they ever move together on one platform that will be something like Win8 that supports both and be still 3-5 years in the future.
Never say Never!!!
a) ....Or just drop support for them without explanation (except to say that we don't need it) as they did with rosetta, adobe flash, and spaces implementations, and just let people suck it up (Apple has never been one to really bend to the demands of the customer. It was always, this is it, take it or leave it. Maybe with Steve gone this attitude has changed some??).
b) The performance may be less, but if it is a chip of their design and manufacturer it could possibly lower costs and thus make them more affordable (thus putting them into a newer category, as they did with the iPad, a place where performance didn't matter in the beginning as there was nothing to measure against. You are assuming as well that they have just developed this processor within the last year. Do you think that Intel just recently developed the IvyBridge? It has been in the works for years in a lab somewhere, as most technology making it to current market is 4-5 or greater years old when it hits the market , just look at Thunderbolt/ Lightspeed. Apple hasn't always been about performance, in the early days they sacrificed performance for the slower chip and lower price of Motorola chips so they have a history of making choices on a cost basis. As to the compiler whizzes at Intel, they said the same thing about early search engine pioneers, until the Google engineers appeared out of nowhere and optimized their functionality like no one dreamed was possible.
As for the merging of OS/IOS, once again we are not sure that this has not already been in the works for the last 5 years (the roadmap set) and Apple is just now on the first leg of actual implementation.
I know lots of people that have gone away from their desktops to lower powered laptops and are just waiting for the day in the near future where they can move entirely to their iPad. These same people are entirely in the cloud as we speak. With cloud based processing (my company is currently rolling out a cloud based sharing system where one can purchase time on a cloud based network for the short times that they need the processing power increases). WIth such development as this, the need for high power portable processing is greatly reduced, and a higher powered A5/A6 chip logging into a powered server farm somewhere would be sufficient for the majority of users needs. Heck even parallels and VMWare have applications that allow one to log into a remote system for the heavier processing power needed. Imagine this on a larger scale implementation, and you can see the reduced need for high powered laptops, and many people will sacrifice power for battery life and weight, hence the success of the iPad and Air.