Animals benefit from phenotypic plasticity in changing environments, but this can come at a cost. Colour change, used for camouflage, communication, thermoregulation and UV protection, represents one of the most common plastic traits in nature and is categorised as morphological or physiological depending on the mechanism and speed of the change. Colour change has been assumed to carry physiological costs, but current knowledge has not advanced beyond this basic assumption. The costs of changing colour will shape the evolution of colour change in animals, yet no coherent research has been conducted in this area, leaving a gap in our understanding. Therefore, in this Review, we examine the direct and indirect evidence of the physiological cost of colour change from the cellular to the population level, in animals that utilise chromatophores in colour change. Our Review concludes that the physiological costs result from either one or a combination of the processes of (i) production, (ii) translocation and (iii) maintenance of pigments within the colour-containing cells (chromatophores). In addition, both types of colour change (morphological and physiological) pose costs as they require energy for hormone production and neural signalling. Moreover, our Review upholds the hypothesis that, if repetitively used, rapid colour change (i.e. seconds–minutes) is more costly than slow colour change (days–weeks) given that rapidly colour-changing animals show mitigations, such as avoiding colour change when possible. We discuss the potential implications of this cost on colour change, behaviour and evolution of colour-changing animals, generating testable hypotheses and emphasising the need for future work to address this gap.