The thing is, there's less air for the blades to move, but there's also less drag on the blades for the same reason, so they spin faster with the same power input. This ends up cancelling out the loss of lift, meaning that the performance stays roughly constant with altitude until it is high enough that the blade tips start to approach the speed of sound.
> In normal operations, and design aims to achieve this, the rotor tips do not go supersonic since when they do, there is a sudden and large decrease in performance with more power required, higher blade loads, vibration and noise.
> Think about a helicopter flying forwards. The advancing blade at its most perpendicular position experiences a relative airflow which is equal (ignoring all kinds of minor side effects) to the forward speed plus the speed of the blade. The retreating blade is experiencing a relative airflow equal to the speed of the blade minus the speed of the helicopter.
> If the blades rotate so fast that the tips are supersonic, then the main lift generating part of the retreating blade, the outer two thirds of the span, would experience such a low airspeed, for some of the span it will even be negative, that the blades will stall causing a catastrophic roll into that side. It is this phenomenon which ultimately limits the rotational speed of the blades and the maximum speed of the helicopter.
