Then they can kiss that backwards compatiability goodbye.....
Not at all. While the original PS3 design was exotic at the time, using an evolution of it after it moves 50Million+ units isn't the same. Sony owns the Cell IP, licensing a GPU from nVidia gives them BC and reduces engineering costs by a huge amount, if they outsorce chip production instead of building a new fab the initial start up costs for the PS4 would be a small fraction of the cost they faced when working on the PS3. It seems the logical choice for Sony is to go with a higher power Cell+nV GPU for the PS4(along with likely ~8GB of RAM, 8-16x BRD and a TBish HD). What CPU MS is going to move to should be interesting, perhaps a Cell variant? Should be cheaper then paying to get an updated version of their current CPU, going with another ATi GPU w/eDRAM just makes sense for them from both a cost to develop angle or a BC one.
Not sure where you got your numbers from, but wiki shows the GC @ 21.7 million and the XBox at 22 million
My numbers show the Xbox at just a but under 21Million units.
And while wild success of their hand held units may have been (and still is) equally undeniable, it wouldn't make sense to keep a declining home console business afloat based on the success of the hand helds.
Their console division was still making a profit, a rather large one too. Unlike most of the other systems, Nintendo not only makes money on the hardware, they make tons of money on the software. If we go back to 2005, the year before the Wii launched, Nintendo sold 27.2Million units of software, MS games studio less then 4Million. Nintendo was extremely profitable prior to the Wii- both its' hardware and software had made them extremely large sums of money. Why would any company consider leaving a market they were making huge gains in when it came to profits? It makes almost no sense.
Also anyone who thinks Nintendo's next handheld will be a Tegra platform is out of their mind.
Tegra is off the shelf and quite cheap, how would it be a bad solution? In the handheld space as of right now, it seems to me that Tegra is far beyond anything else by a healthy margin.
It takes about two years to make a console doesn't it?
Depends on the complexity, PS3 took about twice that long to develop and they still never met the initial design goals(which they would have failed with, Cell is a killer vector processor, but not strong enough to best a GPU).
I think there is a trend within the industry away from loss leading consoles.
Huh? The PS3 and 360 set new records for loss leading, by rather huge amounts too. The Wii isn't a loss leader, but Nintendo has never really used that approach.
So I expect a part that will not be revolutionary but will be very derivative of similar parts from the PC world in order to reduce costs. Something like a 5770 + 2generations + edram. The cpu will probably be a regular quadcore.
They won't downgrade their CPUs, not a chance. The PC space is a bad place to get a CPU for a gaming console, very poor actually. The costs are actually too high and the performance versus what is needed is far too low. This is why every succesful console ever has used something other then a x86 based chip(no, I don't consider the original XBox to be remotely successful).
Fairly safe bet it will be at least 8GB of RAM. PCs are actually falling behind in terms of the size of games when compared to console exclusives(from an assets perspective). The amount of content the larger budgets allow create a lot of oppurtunity for more complex assets.