Non-structural carbohydrates (NSCs) are critical to maintain plant metabolism under stressful environmental conditions, but we do not fully understand how NSC allocation and utilization from storage varies with stress. While it has become established that storage allocation is unlikely to be a mere overflow process, very little empirical evidence has been produced to support this view, at least not for trees. Here we present the results of an intensively monitored experimental manipulation of whole-tree carbon (C) balance (young Picea abies (L.) H Karst.) using reduced atmospheric [CO2] and drought to reduce C sources. We measured specific C storage pools (glucose, fructose, sucrose, starch) over 21 weeks and converted concentration measurement into fluxes into and out of the storage pool. Continuous labeling ((13)C) allowed us to track C allocation to biomass and non-structural C pools. Net C fluxes into the storage pool occurred mainly when the C balance was positive. Storage pools increased during periods of positive C gain and were reduced under negative C gain. (13)C data showed that C was allocated to storage pools independent of the net flux and even under severe C limitation. Allocation to below-ground tissues was strongest in control trees followed by trees experiencing drought followed by those grown under low [CO2]. Our data suggest that NSC storage has, under the conditions of our experimental manipulation (e.g., strong progressive drought, no above-ground growth), a high allocation priority and cannot be considered an overflow process. While these results also suggest active storage allocation, definitive proof of active plant control of storage in woody plants requires studies involving molecular tools.