In a changing climate, organisms are experiencing increasing thermal stochasticity. Despite growing recognition of thermal variability's biological significance, the specific role of environmental unpredictability on organismal performance remains poorly explored. We addressed this gap by testing the hypothesis that stochastic regimes impose higher costs on fish performance beyond those attributable to thermal variance alone. In this context, we investigated the effect of patterns of stochastic and cyclic thermal variations on energetic responses in a freshwater fish, the European chub (Squalius cephalus). Fish (n=130) were exposed to constant temperatures regimes (13°C, 19°C or 25°C) or experienced variable profiles sharing the same mean (19°C) and variance. Variable groups underwent either stochastic variations randomly oscillating between 13°C and 25°C, or cyclic variations following a circadian pattern. Both energetic cost through whole-organism oxygen consumption and energy intake with predation rate were assessed on the same fish after one and six weeks of exposure. Additionally, we measured growth, thermal tolerance, mitochondrial respiration, oxidative stress and telomers length. Despite having greater food intake, fish from the stochastic treatment had similar energetic needs and performance and they grew less than constant average fish, a pattern not visible in cyclic fish. Yet, stochastic fish exhibited similar levels of thermal tolerance, mitochondrial respiration and oxidative stress. These findings point out profound consequences of stochasticity on fish energetic budget, which may be harsher than predictable fluctuations. They suggest an unresolved energetic trade-off and trait specific plasticity to variability, highlighting the urgency to better incorporate stochastic patterns into experimental designs.