Microsoft chose PowerPC because they could get full control of the chip (and for instance embed security features inside it... making it much harder to crack the console). That was not possible with intel or AMD. Original Xbox was cracked almost instantly and it meant major losses due to piracy.
You're forgetting something: You can't just hack something like this (computer HW business introducing a new platform) up from our garage anymore. It takes capital, lots of it. Back in the day it was possible to (literally!) start your business from a garage... I argue that this isn't possible anymore.
How would you go on and sell this idea to VCs? New game console? New desktop? What would bring them the profits for their investment?
And btw, game consoles are cheap because they're sold at a loss. The income comes from the game sales.
Microsoft went to IBM and asked for a CPU within a price range with a minimum performance that's all, which they decided they could provide thanks to CELL processor development work.
This was all confirmed from a contact who deals directly with Microsoft on a daily basis and specifically Xbox360 for licensing things like the visualizer and his 2 games released as full price titles I believe him wholeheartedly when he explains Microsoft went to IBM and IBM used half the research into CELL already subsidised by Sony/Toshiba to create Xenon. This is not based on Wikipedia bollox.
Xenon is 100% owned by IBM due to complaints and legal threats from Sony/Toshiba too.
So you have a situation exactly the same as in 1988 with Amiga A500 vs Sega Genesis/Megadrive. Both machines use the same CPU, and not some crap like the 65816 from WDC like in the pathetically slow SNES, and one costs 2.5x more than the other. The Megadrive was subsidised with £50 games too. The 360 motherboard may well be subsidised but it is still a cheap nasty device which is why it is $200 console not $450 like PS3. Even without subsidy the motherboard essentials are still going to come in at under $250...clearly a massive improvement on $1200 for a GPU devoid weird G5 based mobo used by x1000
The more things change the more they stay the same, my comment was merely if someone was serious about creating a true next gen OS4 platform then the only viable option is to use the 3.2Ghz Xenon CPU and to 'investigate' how Microsoft harness the power, the full price of which is still under $250 which is significantly cheaper than i7 solutions in x86-64 world. Creating the GLUE logic to make an ATI GPU and Xenon and RAM work is not costly or difficult and like I said the option is always there to hire IBM consultants to help as and when required.
It is VERY possible for a new spiritual successor of Amiga to exist, but it takes talent, insight, a rewrite of OS4 and substantially huge balls of solid iron to venture into such a business. Would not be expensive as Xenon is significantly cheaper than i7 due to the rate at which IBM has been producing them for the last 4 years either. Not the A1000 though, like I said more of the 1988 Sega Megadrive vs Amiga A500 kind of level of price/performance/compromise level. Had such design choices been made you could easily have got 1080p gaming rig performance for less than i7 PC gaming rig prices. (Apple are always overpriced and no games for it so forget those).
PC-AT/XT vs A500 vs Sega 16bit console is no different to xbox360 type tech on a desktop machine vs xbox360 vs PC x86-64. Where is the problem? None except for lack of vision. Never going to happen I agree, but the reason why a seriously powerful box to run OS4 on one of its cores at 3.2ghz doesn't exist isn't really a technical issue. Xenon is dirt cheap, ATI GPUs cost peanuts in bulk, basic engineering skill to stick it all together on a custom mobo are plentiful amongst talented graduates etc.
And as I stated before, x1000 is fine if all people want is the fastest OS4 box money can buy, not great value for money and OS4 has its problems too hence my 'pass' on a potential purchase.