I think we could at least draw some inference about this from the Powerbook DL SuperDrive, which uses a variation on SpeedStep to either divide-by-two the processor speed (a feature that goes back a couple of years) or divide-by-four the processor speed (new on the Powerbook DL, optional for the 7447 and required by the 7448, hence only the Powerbook DL can take DayStar's 7448 upgrade).
Bottom line is that Powerbook advertised maximum light use battery life (i.e. no Bluetooth, no Airport, no Photoshop, dim monitor, hard drive on most conservative energy setting) went up from 4.5 hours to 5.5 hours -- but heavy use battery life remained in the 2.5 to 3.0 hours range. Apple acknowledges this in their detailed technical overview document on the Powerbook DL
If it is possible for MacBook Pro users to get up from the 3.0 hour range that everyone seems to be in with all the toys running except the DVD, up towards five hours for light use, that would imply to me that SpeedStep is operational. Remember that SpeedStep does not power down the computer as far as PPC's divide-by-four, but that it is more fine tuned and works in 200MHz increments. A previous poster noted 800MHz as the minimum for the Pentium-M and I seem to recall 1GHz for the Core Duo. Also remember that we already know from Anandtech and other geek sites that the range of battery life on a CoreDuo system seems to be less between light and heavy use than on a Pentium-M.
