For the remote Sahara, the Earth's largest dust source, there has always been a near-absence of data for evaluating models. Here, new observations from the Fennec project are used along with Sahelian data from the African Monsoon Multidisciplinary Analysis (AMMA) to give an unprecedented evaluation of dust-generating winds in the European Centre for Medium-Range Weather Forecasts ERA-Interim reanalysis (ERA-I). Consistent with past studies, near-surface, high-speed winds are lacking in ERA-I and the diurnal variability is under-represented. During the summer monsoon season, correlations of ERA-I with observed wind-speed are low (∼0.35 in Sahel and 0.25–0.4 in the Sahara). Fennec data show for the first time that: (1) correlations are reduced even in the Sahara, not directly influenced by the monsoon, (2) the systematic underestimation of observed winds by ERA-I in the summertime Sahel extends into the central Sahara: potentially explaining the failure of global models to capture the observed global dust maximum that occurs over the summertime Sahara (such as CMIP5), and demonstrates that modelled winds must be improved if they are to capture this key feature of the climatology.