The pattern of neural, physiological and behavioral effects induced by cocaine is consistent with metabolic neural activation, yet direct attempts to evaluate central metabolic effects of this drug have produced controversial results. Here, we used enzyme-based glucose sensors coupled with high-speed amperometry in freely moving rats to examine how intravenous cocaine at a behaviorally active dose affects extracellular glucose levels in the nucleus accumbens (NAc), a critical structure within the motivation-reinforcement circuit. In drug-naive rats, cocaine induced a bimodal increase in glucose, with the first, ultra-fast phasic rise appearing during the injection (latency 6–8 s; ~50 μM or ~5% of baseline) followed by a larger, more prolonged tonic elevation (~100 μM or 10% of baseline, peak ~15 min). While the rapid, phasic component of the glucose response remained stable following subsequent cocaine injections, the tonic component progressively decreased. Cocaine-methiodide, cocaine's peripherally acting analog, induced an equally rapid and strong initial glucose rise, indicating cocaine's action on peripheral neural substrates as its cause. However, this analog did not induce increases in either locomotion or tonic glucose, suggesting direct central mediation of these cocaine effects. Under systemic pharmacological blockade of dopamine transmission, both phasic and tonic components of the cocaine-induced glucose response were only slightly reduced, suggesting a significant role of non-dopamine mechanisms in cocaine-induced accumbal glucose influx. Hence, intravenous cocaine induces rapid, strong inflow of glucose into NAc extracellular space by involving both peripheral and central, non-dopamine drug actions, thus preventing a possible deficit resulting from enhanced glucose use by brain cells.