Made some measurements powering a A2212 1000KV motor with a 1045R propeller from a 30A ESC powered with 3S LiPo battery, the ESC was driven with servo inputs 1000-2000 as the power input to the ESC and thrust generated by propeller were measured. The attached graph with PWM width x axis shows the power going into the ESC with varying throttle settings 1000-2000.
The data was rearranged a bit and graphed on the second plot showing thrust as a function of the motor input power (which was normalized by subtracting out the no propeller power measurement at each throttle setting). A curve was fit to the data showing
the thrust out/power in efficiency decreasing as the 0.65 power. Maybe the drag resistance increases at close to second power of angular speed while thrust increases linearly with angular speed?
Playing with the numbers a bit it looks like my ~1Kg F450 will use about 75watts to hover with the four motors each contributing 1/4Kg of thrust.