It is known that the total power consumption of a device is determined by the sum of the power of the nodes. And while the power of each node will be determined by the supply voltage and current consumption of this node. In a word, I think that if you do not take into account the power part of the PI controller, then taking into account the current consumption of microcircuits and controller cascades in the usual way, multiplying the current consumption by the voltage value, we find the power consumed (lost) on the controller.