Competition among plants in extreme environments such as the High Arctic has often been described as unimportant, or even nonexistent; environmental factors are thought to overrule any negative plant–plant interactions. However, few studies have actually addressed this question experimentally in the Arctic, and those that did found only little evidence for competition. Such species interactions will presumably become more important in the future, as Global Climate Change takes effect on terrestrial ecosystems. We investigated plant–plant interactions in the High Arctic, following the growth of Luzula confusa and Salix polaris in pure and mixed stands, and under elevated-temperature treatment over 2 years. To understand the mechanisms of competition, a parallel experiment was undertaken in phytotrons, manipulating competition, temperature and nutrient availability. Our findings indicate that competition is acting in the natural vegetation, and that climatic warming will alter the balance of interactions in favour of the dwarf shrub S. polaris. The phytotron experiment suggested that the mechanism is a higher responsiveness of Salix to nutrient availability, which increased under warming in the field. While Luzula showed a positive response to higher temperature in the lab, its performance in mixed stands in the field was actually reduced by warming, indicating a competitive repression of growth by Salix. The growth of Salix was also reduced by the presence of Luzula, but it was still able to profit from warming. Our findings suggest that climatic warming will result in greater shrub dominance of High Arctic tundra, but we also conjecture that grazing could reverse the situation to a graminoid-dominated tundra. These two divergent scenarios would have different implications for ecosystem feedbacks to climatic change.