My guess is the rumour is true but the analysis is misguided. I can't see Apple making this enormous change simply for GHz reasons. IBM seems able to deliver both low power and extremely high-performance chips as we have seen with the Xbox and Cell (PS3) processors. X86 is probably a better long term bet, but I doubt the need switch now just for this reason.
I think the main motivation is to gradually position the Mac as a "Better PC", which is able to run Windows software while delivering a better user experience than a PC. Apple needs to do this to secure access to Office, which is much less guaranteed than the supply of PPC chips now that Apple openly competes with Office. Also, Apple needs to reach that 95% of users for whom 1% of Windows software is critical.
I'd expect Apple to make a deal with VMware to offer excellent PC emulation, or even pull off some Wine-like system that lets users mix Win32 and OSX apps. They have done this quite well with the OS9 virtual environment. Apple can easily port the Finder, OS utilities, music and productivity sofware that the bulk of the users use, without having to lean on any third party developers. They could even make the switch free and automatic via the Setup Assistant (a built-in tool for migrating from an old to a new Mac).
Users would be delighted to have the superior usability and aesthetics of a Mac for most of their user experience (finding files, mail, web, etc) as well as the ability to run Windows apps, albeit with the slight shortcoming that they look and behave like Windows apps. Delivering all this at a modest price (I'd expect a small premium for great industrial design) seems like a very good proposition to consumers.
As for Apple software developers, a switch would allow Apple to deliver tools that let ISVs write (nice looking) Apple apps that run on Windows, like iTunes. Access to the Windows market in this way ought to be a very welcome prospect to developers like Omni.