[aerohead]

I've been thinking about it since you first posted it. Some random thoughts:
* It's very easy to spin an unloaded fan.
* The mechanical advantage of any induced aerodynamic force on the blades, acting away from the shaft center, can easily overcome any initial bushing / bearing resistance to movement.
* Once spinning, it requires very little energy to maintain the rotation.
* After all, they're not harvesting any power from the airstream to produce any useful work.
--------------------------------------------------------------------------------------
* The fan is extremely 'porous'.
* A lot of air bleeds through the unit, whether through the blades or not.
--------------------------------------------------------------------------------------
* The air which makes it through may be deflected in direction a bit, or blasted into turbulence, if it wasn't turbulent to begin with.
* As an additional source of resistance, any chaotic, random turbulent air it's initially embedded within, may end up a little less energetic due to viscous shearing effects, however I don't know that they'd be significant enough to even measure.
* Conservation of energy dictates that whatever kinetic energy existed in the air before it encountered the fan would be preserved, minus a minute amount lost to atmospheric heating due to the viscous shear.
* It couldn't amount to too much. We're adding a smidgeon of additional disorder to a completely turbulent/ chaotic environment.