Treelines in high latitudes and high altitudes are considered sentinels of global change. This manifests in accelerated encroachment of trees and shrubs and enhanced plant productivity, with currently unknown implications for the carbon balance of these biomes. Given the large soil organic carbon stocks in many treeline soils, we here wondered whether introducing highly productive plants would accelerate carbon cycling through rhizosphere priming effects and if certain soils would be more vulnerable to carbon loss from positive priming than others. To test this, organic and mineral soils were sampled above and below treelines in the Swedish sub-arctic and the Peruvian Andes. A greenhouse experiment was then performed to quantify plant-induced changes in soil mineralisation rates (rhizosphere priming effect) and new C formation using natural abundance labelling and the C4- species Cynodon dactylon. Several environmental, plant, soil and microbial parameter were monitored during the experiment to complement the observations on soil C cycling. Priming was predominantly positive at the beginning of the experiment, then systematically decreased in all soils during the plant growth season to be mostly negative at the end of the experiment at plant senescence. Independent of direction of priming, the magnitude of priming was always greater in organic than in corresponding mineral soils, which was best explained by the higher C contents of these soils. Integrated over the entire study period, the overall impact of priming (positive and negative) on the soil C balance was mostly negligible. Though, net soil C loss was observed in organic soils from the sub-arctic tundra in Sweden. Most notably, positive and negative priming effects were not mutually exclusive, rather omnipresent across ecosystems, depending on sampling time. The direction of priming seems to be fluctuating with plant productivity, rhizosphere carbon inputs and nutrient uptake. This highlights the need for integrative long-term studies if we aim to understand priming effects at ecosystem scale and greenhouse and laboratory studies must be validated in situ to enable reliable ecological upscaling.