Reforestation and afforestation often takes place in the form of monocultures. These have increasingly come under criticism due to their greater susceptibility to adverse environmental conditions and pathogens and because of the negative long-term impacts they have on soil fertility. Moreover, reservations have frequently been expressed as to whether monoculture is the only way to maximize biomass productivity. Subtropical forests in China are not only rich in tree species but are also diverse in terms of topography. However, historical clear-cutting and the high demand for wood have meant that considerable areas of these forests have been degraded or lost entirely. In recent years, great efforts have been made in subtropical China to restore and afforest vast areas. Nonetheless, conifer monoculture plantations are still prevalent owing to the lack of data on the relationship between diversity and ecosystem functioning and the limited knowledge of the growth performance of native broad-leaved tree species in plantations in this region. In this study, we investigated the growth response of individual tree saplings in a large-scale forest biodiversity and ecosystem functioning (BEF) experiment established in a highly heterogeneous environment in subtropical China (BEF-China). The experiment was established during the period November 2008 to April 2009 and incorporates a species pool of 24 tree species across 261 plots (0.067 ha each) and a broad richness gradient with 1–24 species per plot. Annual growth rates of tree height, stem diameter and crown width for 6723 saplings of 24 species in the third year after establishment were used as response variables. We analyzed the effects of initial size, local abiotic site conditions and local neighborhood on growth performance using mixed effects models. Initial size, local topographic (aspect, slope, curvature, elevation) and edaphic variables (total nitrogen, C/N ratio) as well as local neighborhood variables (competition index, Shannon index) were included as fixed effects. Random effects were plot and species identity. The results showed that growth performance of tree saplings was markedly species-specific and strongly affected by initial size and local site conditions. In this early stage of the experiment, local Shannon diversity did not yet explain a significant amount of variation in sapling growth. However, the significant effect of the increasing size of neighboring trees indicated that tree–tree interactions had already occurred. A greater investment in height growth at the expense of diameter growth suggested that light intercept is the priority for tree sapling growth. This study stresses the importance of better understanding the effects of local site conditions on initial growth in mixed-species plantations.