Cognitive radio and other dynamic spectrum sharing paradigms have enjoyed the limelight recently for their potential to realize a number of benefits for communications systems. These benefits have included the achievement of greater spectral efficiency, and the ability to offer service provision and/or improved QoS to a range of networks and devices. However, one opportunity that has not been significantly investigated is the potential for dynamic spectrum sharing to reduce necessary transmission powers in wireless communications, and to reduce necessary power draw in running network equipment per se. This paper investigates such potential through four means: spectrum sharing to enable the movement of users between bands allowing radio network equipment in other bands to be switched off, better interference management to reduce necessary transmission power, spectrum sharing to improve links’ propagation characteristics therefore reducing necessary transmission power, and the aggregation of spectrum, increasing channel bandwidths, therefore reducing the necessary transmission power along the lines of Shannon and other various capacity formulas. Power consumption reduction potential of over 50%, both in terms of “from-the-socket” power and in terms of necessary transmission power, is demonstrated.