Abstract Improved understanding of hydraulic fractures is needed to optimize petroleum well drilling and completion strategies. Yet direct observations of hydraulic fractures are rarely made, and reliance is placed on indirect methods such as microseismic monitoring, interference tests, fibre optic detection of fracturing in adjacent wells and numerical modelling. While these techniques provide useful insights, verification of such studies is commonly lacking; core taken through stimulated intervals offers a robust option for verification. We make the case that core can provide complementary information at a different scale from other data types. Core from a slant well adjacent to two stimulated wells at the Hydraulic Fracture Test Site (HFTS1) in West Texas revealed 375 hydraulic fractures, striking 090°±20°, subparallel to present-day S Hmax . There are more hydraulic fractures than expected, and clustering across a range of scales from approximately 1 cm to 50 m. The largest cluster correlates with high microseismic event density. Diversion, bifurcation, and segmentation structures may account for the large number of fractures observed and the orientation spread. Reactivation of sealed natural fractures and bedding planes is relatively uncommon. Proppant sand packs and patches occur in a few locations, particularly where hydraulic fractures are complex.