Experimental evidence has demonstrated that glutamate is an essential factor for neurogenesis, whereas another line of research postulates that excessive glutamatergic neurotransmission is associated with the pathogenesis of depression. The present review shows that such paradox can be explained within the framework of hormesis, defined as biphasic dose responses. Low glutamate levels activate adaptive stress responses that include proteins that protect neurons against more severe stress. Conversely, abnormally high levels of glutamate, resulting from increased release and/or decreased removal, cause neuronal atrophy and depression. The dysregulation of the glutamatergic transmission in depression could be underlined by several factors including a decreased inhibition (γ-aminobutyric acid or serotonin) or an increased excitation (primarily within the glutamatergic system). Experimental evidence shows that the activation of N-methyl-D-aspartate receptor (NMDA) and α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors (AMPAR) can exert two opposite effects on neurogenesis and neuron survival depending on the synaptic or extrasynaptic concentration. Chronic stress, which usually underlies experimental and clinical depression, enhances glutamate release. This overactivates NMDA receptors (NMDAR) and consequently impairs AMPAR activity. Various studies show that treatment with antidepressants decreases plasma glutamate levels in depressed individuals and regulates glutamate receptors by reducing NMDAR function by decreasing the expression of its subunits and by potentiating AMPAR-mediated transmission. Additionally, it has been shown that chronic treatment with antidepressants having divergent mechanisms of action (including tricyclics, selective serotonin reuptake inhibitors, and ketamine) markedly reduced depolarization-evoked glutamate release in the hippocampus. These data, taken together, suggest that the glutamatergic system could be a final common pathway for antidepressant treatments.