One additional consideration in favor of a move in this direction is that Microsoft's Surface Computer has been getting good reviews, and that one MS executive stated that there was a $10 billion market for the device. (E.g., as ordering devices in restaurants, as gaming devices in retail spaces and airlines, as informational kiosks everywhere, etc.) Apple mightn't want to let MS establish too great a lead in buyers' mindspace in such a rich market.
OTOH, if Apple did come out with built-in touch in its iMac line, the only way this would lock out competitors would be if independent developers were encouraged to rely on the touch interface. But why would they do that, if it would lock out 90% of the Mac user base? Another OTOH is that it seems unlikely that a rumor of such a wide-ranging hardware transition hadn't appeared earlier.
Anyway, if touch is the new transition, then I hope Apple will avoid a safe, beige, "left-brained" name like "Surface Computer." I suggest TapTop. It's more memorable, being "anchored" in laptop. It's almost unforgettable. And it has good overtones, suggesting things like "tiptop." ("Surface" has bad overtones, suggesting "shallow," "brittle," etc.)
Transition candidates would be more plausible if they involved an innovation that could be kept a secret until just before release. One such innovation would be a chip that would give a Mac built-in, hardware-based "software metering," so that a user would be able to run software on a rental basis. This would give users inexpensive access to expensive software they would only rarely use--but that would be quite rewarding to them anyway. They'd have access to much more software than on any competing platform. It would be a tremendous selling point--and only a company that controlled the hardware could offer this in a way that was secure enough to reassure software vendors. (E.g., perhaps there would be online monitoring by Apple of metered computers to ensure that the security of the metering hadn't been compromised.)
It would appeal to software developers because it would lower their marketing expense--and probably reduce piracy considerably as well. Small software companies challenging established giants would particularly like this, because they wouldn't have to get customers to shell out big bucks to displace their competitors.
Even if this isn't what Apple has in mind, it ought to be. (Maybe next year.)
Alternatively, maybe it's a built-in chip that performs speedy encryption, or that monitors the computer for malware. (Again, if it isn't, it should be--eventually.)