In this study, we have demonstrated that using dry matter intake (DMI) phenotypes from multiplecountries increases the accuracy of genomic breeding values for this important trait, provided a multi-trait approach is used. Data from Australia, Canada, Denmark, Germany, Ireland, the Netherlands,New Zealand, United Kingdom and two institutions in the United States were combined to estimatethe accuracy of genomic prediction for DMI multi-trait models. The average accuracies was 0.44, andranged from 0.37 (Denmark) to 0.54 (the Netherlands). Enlarging the combined dataset with uniquephenotypes does increase the accuracy of the genomic prediction for DMI. This stimulates furtherinternational collaboration.