cAMP-dependent protein kinase (PKA) plays a central role in important biological processes including synaptic plasticity and sympathetic stimulation of the heart. Elevations of cAMP trigger release of PKA catalytic (C) subunits from PKA holoenzymes, thereby coupling cAMP to protein phosphorylation. Uncontrolled C subunit activity, such as occurs in genetic disorders in which regulatory subunits are depleted, is pathological. Anchoring proteins that associate with PKA regulatory subunits are important for localising PKA activity in cells. However, anchoring does not directly explain how unrestrained ‘free swimming' of C subunits is avoided following C subunit release. In this review, I discuss new mechanisms that have been posited to account for this old problem. One straightforward explanation is that cAMP does not trigger C subunit dissociation but instead activates intact PKA holoenzymes whose activity is restrained through anchoring. A comprehensive comparison of observations for and against cAMP-activation of intact PKA holoenzymes does not lend credence to this mechanism. Recent measurements have revealed that PKA regulatory subunits are expressed at very high concentrations, and in large molar excess relative to C subunits. I discuss the implications of these skewed PKA subunit concentrations, before considering how phosphorylation of type II regulatory subunits and myristylation of C subunits are likely to contribute to controlling C subunit diffusion and recapture in cells. Finally, I speculate on future research directions that may be pursued on the basis of these emerging mechanisms.