well, for one the PowerPC 970 won't 'save us', as it's still almost a year away, and while its specs would be cool for MWSF PowerMacs, they will be outdated when the processor finally arrives. Then again, we've been living against Moore's Law for quite some time on the Mac platform, so why care...
I was mentioning market share only because Apple today makes its money in hardware. If they would want to switch to being a software only company (I don't say they _do_ want that, just if), they'd have to gain some market share to make enough money to survive - or cut down most of their financial plans, which of course isn't much of a strategy...
About your examples, RacerX...
sgi: They tried to go Intel, yes. But they didn't only fail because of a lack of developer support. They had a different problem altogether, they are in a dying market. With 3D cards coming out for GAMING every two months that put their expensive hardware to dust, they didn't have much of a market. I still don't know why exactly they went Intel, but it just seems (looking back) like they just had to do _something_.
Sun: Never was in the market for a desktop system (well, maybe now with their Linux effort, but not then...)
Apple: They were a dying company when the Rhapsody plans were first announced. And remember why Mac OS X (with Carbon) came up? Because the developers wouldn't want to port to Cocoa - period. Not much to do with X86 or not, although I agree that there was more interest in the PPC version back then.
I'm only trying to figure out what would/could/should happen if those rumours are in fact true and Apple plans to do that step and release Voyager as a software product (operating system, iApps, their apps...).
And my personal guess would be that companies like Adobe would basically welcome the move, if they could really just recompile their Carbon apps. (But that's not exactly defined, isn't it...)
But what about viral marketing? Apple could release Voyager with only Cocoa support plus the iApps and their apps (Final Cut Pro mainly, DVD Studio Pro maybe...), along with the development tools and Darwin's source code. And before long, we would see some development that could also help the Mac (PPC) platform, because Apple would of course make sure that apps developed on Voyager would easily compile for the PPC platform, too.
Heck, *if* Apple has an almost ready version of Voyager and they _don't_ think developers will jump on _that_ bandwaggon, they could even release the baby for free, because, as you say, no apps for the platform and no developer interest would not harm Macintosh sales much. But if you were a PC user and got OS X for free, you'd certainly see how good it was, and because the apps would only be out for PowerPC versions of X, you would maybe buy an iBook or two (j/k about the 2).
My main point however is that Microsoft is losing ground. Although nobody ever really liked them, they were sometimes respected as a good developer for business operating systems. Businesses choose Microsoft. They have things like site-licenses. They're not cheap, but if a business asks Apple about something like that, an Apple rep (just my impression) would likely say something like: "No, but you get a free iPod if you order 10 iMacs." Well, Microsoft is losing image. Again and again. And at some point, the time has come to make another attack. Linux will do it anyway. See like RedHat, SuSE, IBM, Sun, Xandros, Lindows and basically just everyone in the _business_ of Linux is pushing Linux to the corporate desktop? And they have a _case_! If you have to support 300 desktops with Office, ZERO is a nice figure for the price of operating system plus office applications. At least compared to what Microsoft is offering. Of course, other stuff costs money, too. But budgets are tight these years. We're not in the 'I need a website' age any longer. Linux has a case and will take a piece of that cake. And I sure hope that Apple will be ready to have some, too.
(Wow, sorry for the somewhat longish post...)