Recent studies have shown that cannabis may disrupt glutamate (Glu) signaling depressing Glu tone in frequent users. Current evidence have also consistently reported lower Glu-levels in various brain regions, particularly in the medial prefrontal cortex (mPFC) of chronic schizophrenia patients, while findings in early psychosis (EP) are not conclusive. Since cannabis may alter Glu synaptic plasticity and its use is a known risk factor for psychosis, studies focusing on Glu signaling in EP with or without a concomitant cannabis-usage seem crucial. We investigate the effect of cannabis use on prefrontal Glu-levels in EP users vs. both EP non-users and healthy controls (HC). Magnetic resonance spectroscopy was used to measure [Glu mPFC ] of 35 EP subjects (18 of whom were cannabis users) and 33 HC. For correlative analysis, neuropsychological performances were scored by the MATRICS-consensus cognitive battery. [Glu mPFC ] was lower in EP users comparing to both HC and EP non-users (p mPFC ]-decline with age was observed in EP users (r = -.46; p = 0.04), but not in EP non-users or HC. Among neuropsychological outcomes, working memory was the only domain that differentiates patients depending on their cannabis use, with users having poorer performances. Cannabis use is associated with reduced prefrontal [Glu mPFC ] and with a stronger Glu-levels decline with age. Glutamatergic abnormalities might influence the cognitive impairment observed in users and have some relevance for the progression of the disease.