I think most people now will assume any new Nintendo hardware is weak unless shown otherwise and Nintendo nowadays won't even give out performance information so even if Nintendo bought a reasonable AMD chipset to power their next console they will probably fail horribly regarding informing the end consumer about it that will probably damage it's chances anyway.
I think the success of the next console is down to their new gimmick they have come up with. How good is this gimmick though because they are keeping it under wraps so may not have done any consumer research on its appeal. It may end up like the wii u tablet as a huge failure. Their instincts were completely wrong last time.
Actually scratch that I don't think their instincts were wrong I think it was the implementation that was horrible. I still think a normal size controller with a small touch screen would have done well with decent battery life and analogue triggers and the money saved on that could have meant the wii u console was powerful enough to be at least cleanly more powerful than 360 and PS3 to get enhanced versions of their games rather than inferior versions. I think that console would surely have doubled the current sales number and at least been competitive with xbone even if the ps4 was still a runaway success. That wii u would have found it easier to do 720p 60fps games with anti-aliasing simply by perhaps using a couple of 64bit x86 cores at 1.8ghz and have the backwards compatibility ppc at 1.8ghz too, the gpu could have been around 500 gflops all in a 28nm fabrication package with thermal dynamics no worse than the current wii u at 45nm. This design would have spanned both the old generation and new generation for code compatibility.
