I think this could really rescue the GT200 series as viable. If you look at the problems, it's not that it's big or monolithic or whatever, it's that it's not fast enough to justify the size. And if you further boil it down, the achilles heal of GT200 is simple clockspeed. That is why a 9800GTX pushes a GT260, because the shaders are clocked so much higher on the 9800 it has nearly as much shader power. And the rest of the core is clocked higher on the 9800 giving it more texture ability as well. GT260 offers more BW, ROPS, and RAM, but in many scenarios those dont come into play.
So basically if GT200b can address the clockspeed problems (as well as cost problems to some extent), it will really help. I also dont see a way for AMD to easily respond, if say the 200B based 260 does trump the 4870. It's not like AMD should have a new chip anytime soon and they're already at 55nm. Of course AMD should still have plenty of cost advantages though.
So basically if GT200b can address the clockspeed problems (as well as cost problems to some extent), it will really help. I also dont see a way for AMD to easily respond, if say the 200B based 260 does trump the 4870. It's not like AMD should have a new chip anytime soon and they're already at 55nm. Of course AMD should still have plenty of cost advantages though.