This paper studies the fundamental trade-offs between power transfer efficiency (PTE) and spectral efficiency that occur during simultaneous power and data transfer through near-field inductive links. A mathematical analysis is used to establish the relationship between PTE and channel capacity as a function of link parameters such as coupling coefficient ( k), load resistance, and surrounding environment. The analysis predicts that the optimum trade-off between power and data transfer is particularly dependent on k, which is a monotonically-decreasing function of axial distance ( d) between the coils. Real-time adaptation of the link parameters (such as load resistance and modulation type) is proposed to automatically optimize the power-data trade-off over a wide range of distances and coupling coefficients. A bench-top prototype of such an adaptive link is demonstrated at a center frequency of 13.56 MHz. The prototype uses an ultrasound transducer to measure d with accuracy mm, and uses this information to autonomously optimize both data rate (up to ∼ 50 Mbps) and PTE (up to ∼ 25%) as the coil-coil distance varies within the 4-15 mm range.