I can see why it's a hard thing to come up with, but I'm puzzled as to people's problem accepting it.

The blades create backward-thrust, pushing against the wind, accelerating the car, creating more thrust, pushing harder against the wind, etc. It stops a seemingly-endless increase when you hit the break-even point.

The reason it goes faster is because, when it finally catches up to the speed of the wind (which is obviously possible), the blades are already spinning, thrusting the car forward beyond just the wind speed. If they weren't spinning at all, 1x wind speed would be the max, fairly easily approached (though not achieved). This video is particularly telling of this effect: http://www.youtube.com/watch?v=dgHBDESd38M&feature=relat...