Bird collision assessments are generally made at the scale of a single wind farm. While especially in offshore situations such assessments already hold several assumptions, even bigger challenges exist on estimating the cumulative impact of multiple wind farms and the impacts at population level. In this paper, the number of collision victims at Belgian offshore wind farms was estimated with a (theoretical) collision risk model based on technical turbine specifications, bird-related parameters and bird density data of both local seabirds and passerine migrants. Bird density data were gathered by visual censuses and radar registrations. The outcome of the model was extrapolated to future development scenarios in the Belgian part of the North Sea and in the entire North Sea, and then further used for a preliminary assessment of the impact at population level for the species at risk. The results indicate that the cumulative impact of a realistic scenario of 10,000 turbines in the North Sea might have a significant negative effect at population level for lesser and great black-backed gull. We further show that during a single night of intense songbird migration, the number of collision victims among passerine migrants might be in the order of magnitude of several thousands in the entire North Sea. We argue that it is of great importance to further develop methods to quantify the uncertainties and to minimise the assumptions, in order to assure more reliable cumulative impact assessments.