1. A pressing challenge is to determine whether and how global-change drivers influence species physiology and survival. Recently, researchers have proposed the metabolic theory of ecology, defending the hypothesis of a universal ther- mal dependence of metabolic rate or, alternatively, the metabolic cold adaptation theory, stating that local adaptation can influence the thermal sensitivity of metabolic rate. However, the long-term (i.e. multigenerational) consequences of warming for the thermal sensitivity of metabolic rate remain largely unexplored although it determines energy use and is crucial for species response to climate change. 2. In this study, we used an evolutionary experiment with medaka fishes Oryzias latipes maintained for more than 12 generations at warm and cold temperatures (30 and 20°C, respectively) to address this issue. Our objective was to investigate whether thermal adaptation influences the relationship between temperature and mass-corrected metabolic rate and how this may occur. 3. In agreement with the universal thermal dependence hypothesis, we found that warming did not significantly influence the thermal sensitivity of mass-corrected metabolic rate: neither the intercept nor the slope of the temperature–metabolic rate relationship differed among fish lineages. Our small-scale laboratory experiment thus indicated that there is limited potential for evolutionary change in medaka fish metabolic rate in response to warmer temperatures. 4. Overall, we provide evidence that 9 years of experimental warming did not influence the thermal sensitivity of metabolic rate. Our results highlight the invariability of the thermal dependence of metabolic rate, which has important implications for adaptation to climate warming. This finding suggests a limited potential for metabolic adaptations in response to long-term temperature changes, which may have negative consequences for the persistence of fish populations under climate change.