How ecological and evolutionary processes interact and together determine species and community responses to climate change is poorly understood. We studied long-term dynamics (over approximately 200 asexual generations) in two phytoplankton species, a coccolithophore (Emiliania huxleyi), and a diatom (Chaetoceros affinis), to increased CO2 growing alone, or competing with one another in co-occurrence. To allow for rapid evolutionary responses, the experiment started with a standing genetic variation of nine genotypes in each of the species. Under co-occurrence of both species, we observed a dominance shift from C. affinis to E. huxleyi after about 120 generations in both CO2 treatments, but more pronounced under high CO2. Associated with this shift, we only found weak adaptation to high CO2 in the diatom and none in the coccolithophore in terms of species’ growth rates. In addition, no adaptation to interspecific competition could be observed by comparing the single to the two-species treatments in reciprocal assays, regardless of the CO2 treatment. Nevertheless, highly reproducible genotype sorting left only one genotype remaining for each of the species among all treatments. This strong evolutionary selection coincided with the dominance shift from C. affinis to E. huxleyi. Since all other conditions were kept constant over time, the most parsimonious explanation for the dominance shift is that the strong evolutionary selection was driven by the experimental nutrient conditions, and in turn potentially altered competitive ability of the two species. Thus, observed changes in the simplest possible two-species phytoplankton “community” demonstrated that eco-evolutionary interactions can be critical for predicting community responses to climate change in rapidly dividing organisms such as phytoplankton.