Yes that is true, but I do not think "from switching loses". If I remember well 30 years my courses, switching loses are linear current * voltage (beside frequency, switching time, resistance etc ) ... means not only current. Means if the same device (ESC) manages the same power, does not matter if it is 1A and 10V or 10A and 1V, loses will be the same, also duty cycle does not change it.
Switching losses are more complicated than that, and they definitely depend on the voltages involved as well as the currents. I'm always at my most dangerous when I'm trying to be authoritative before I've had any caffeine, but here goes:
Switching loss in a transistor happens because the transistor does not turn off immediately. The loss happens when the transistor is trying to pinch off the flow of current. So it is the voltage and current
at the transistor, times the amount of time it takes for the transistor to turn off, times the switching frequency. Roughly, the current is the current in the motor, at the moment. When the supply voltage is higher than the voltage the motor needs, the current at the "more driving" transistor is higher than the average current from the supply (how much higher depends on various voltages, the switching speed, and the motor inductance). The current depends on supply voltage, the motor's back-emf (i.e., how fast it's turning), the motor inductance, the duty cycle, and the switching frequency. How quickly the transistor turns off depends on the motor inductance, the voltage on the transistor at the instant that turn-off is commanded, and the gate driver circuit.
In general, the switching losses are at their worst when the duty cycle is 50%, and with all things being equal they get worse as the supply voltage goes up. But there are so many details about circuit design, switching frequency, etc., that you can only make the most general of statements about what will affect things.
My knowledge is a bit more recent than 30 years: the second to last circuit board that I designed for a customer, over the past two years or so, is a switched motor driver. It's much like the ESCs that we use except that it's chock full of bells and whistles and there was no driving need to make it super compact, so a board that'll deliver 200W to two motors is about 130mm x 180mm, and even then my layout guy whined about it being tight.