The emission of artificial light at night (ALAN) is rapidly increasing worldwide. Yet, evidence for its detrimental effects on various species is accumulating. While the effects of ALAN on phenotypic traits have been widely investigated, effects on the molecular level are less well understood. Here we aimed to integrate the effects of ALAN at the transcriptomic and the phenotypic level. We tested these effects on Chironomus riparius, a multivoltine, holometabolous midge with high ecological relevance for which genomic resources are available. We performed life-cycle experiments in which we exposed midges to constant light and control conditions for one generation. We observed reduced fertility under ALAN from which we predicted the population size to decline to 1% after 200 days. The transcriptomic analysis revealed expression changes of genes related to circadian rhythmicity, moulting, catabolism and oxidative stress. From the transcriptomic analysis we hypothesised that under ALAN, oxidative stress is increased, and that moulting begins earlier. We were able to confirm both hypotheses in two posthoc experiments, showing that transcriptomics can be a powerful tool for predicting effects on higher level phenotypic traits.