In iStat, I see the same thing as far as power goes. At the moment I'm on the nVidia card, and under Power at the bottom GPU 1 is 0.75 W, GPU 2 is at 1.20 W (under Current and Voltages, though, both are zero.) Only dependant app I'm using is Tweetie, though I've got it hooked to an external monitor.
When I switch to Intel only both GPUs go to 0.00 under Power.
In my observations the average battery life estimate while it's forced to Intel only seems to be roughly an hour faster than doing the same activity with Nvidia on only, or with dynamic switching enabled.
I think as far as battery life goes two things should really be understood:
1) Apple's own estimates say that turning dynamic switching on (as opposed to leaving just the dedicated GPU on) only gives you an extra hour in battery life under their tests.
2) The battery time is an estimate that fluctuates and updates itself depending on what you're doing with the system. When I was messing around earlier, I had the display dimmed down to one bar, wifi off with nothing open but Scrivener and the estimate oscillated between 5 hours and 8 when it was forced to Intel only. This was such a wide margin when I was doing nothing more with the computer except typing.
It's pretty unlikely that we're going to be seeing eight or nine hours of battery life in any real-world situation. Now Apple said that the battery only gains an hour of life when dynamic switching is enabled, so it could be the case that you'd get more out of it when it's set to Intel only, but if anyone's hoping for this tool to magically give them twelve hours of battery life I think they're going to be disappointed.
