Flagellated bacteria, like Escherichia coli, swim by rotating helical flagellar filaments powered by rotary flagellar motors at their base. Motor dynamics are sensitive to the load it drives. It was previously thought that motor load was high when driving filament rotation in free liquid environments. However, torque measurements from swimming bacteria revealed substantially lower values compared to single-motor studies. We addressed this inconsistency through motor resurrection experiments, abruptly attaching a 1-micrometer-diameter bead to the filament to ensure high load. Unexpectedly, we found that the motor works with only half the complement of stator units when driving filament rotation. This suggests that the motor is not under high load during bacterial swimming, which we confirmed by measuring the torque-speed relationship by varying media viscosity. Therefore, the motor operates in an intermediate-load region, adaptively regulating its stator number on the basis of external load conditions. This ensures the robustness of bacterial motility when swimming in diverse load conditions and varying flagella numbers.