In computer-mediated communication, social categories such as race and ethnicity have to be actively performed and constructed by participants in order to gain visibility; it can be argued that new forms of super-diversity and their sociolinguistic implications become particularly tangible here. As a consequence, such racialized discourse provides ideal material for a sociolinguistic analysis of CMC. Based on these assumptions, this study focuses on how race and ethnicity are performed on the web forum Nairaland, a digital community and place of interaction for Nigerian locals, first- and secondgeneration Nigerian emigrants, as well as participants with other ethnic backgrounds. A large-scale corpus (17 million token, time span of 4 years) was analyzed in terms of racial and ethnic identity construction of the community members; in particular, the use of Nigerian Pidgin as an ethnolinguistic repertoire within the community was taken into account. The analysis includes visualizations of the globalized community structure, a quantitative assessment of the distribution of racial and ethnic labels, and a qualitative close reading of diasporic narratives of belonging. The results of this study illustrate how the use and (often conscious) selection of ethnolinguistic repertoires contribute to the complex and varied racial/ethnical identities on display in the forum data. In this sense, this paper makes a contribution to our understanding of the sociolinguistic implications of super-diversity, and the essential role that digital mediation plays in its emergence.