In this work, we propose Dilated Point Convolutions (DPC) which drastically increase the receptive field of convolutions on 3D point clouds. As we show in our experiments, the size of the receptive field is directly related to the performance of dense tasks such as semantic segmentation. We look at different network architectures and mechanisms to increase the receptive field size of point convolutions and propose in particular dilated point convolutions. Importantly, our dilation mechanism can easily be integrated into all existing methods using nearest-neighbor-based point convolutions. To evaluate the resulting network architectures, we visualize the receptive field and report competitive scores on the task of 3D semantic segmentation on the S3DIS and ScanNet datasets.