Humeral shaft nonunions are challenging to treat, and those that require >1 surgical procedure in order for a nonunion to heal are termed recalcitrant. Most studies on nonunion have evaluated the union rate independent of the number of procedures required to achieve union. The aims of the present study were (1) to compare the healing rates after the index operation for the treatment of a nonunion with conventional versus locked plating with or without graft augmentation, (2) to report the prevalence of recalcitrant nonunion, and (3) to identify risk factors that predict a recalcitrant nonunion. We performed a retrospective analysis of a prospectively collected database of 125 humeral shaft nonunions treated with open reduction and plate fixation by a single surgeon over 25 years. Univariate and multivariate regression analyses were performed to compare healing rates by type of plate fixation and biological augmentation and to identify demographic, injury, and treatment-related risk factors for the development of a recalcitrant nonunion. One hundred and five patients (84%) had healing after the index procedure for the treatment of nonunion. Twenty patients (16.0%) required secondary procedures and were defined as having a recalcitrant nonunion. Eight of these patients (6.4% of the overall group) healed after the secondary interventions, and 12 (9.6% of the overall group) had a failure to unite. There were no significant differences in healing rates between conventional and locked plates or between the types of bone graft (autogenous or recombinant human bone morphogenetic protein). Risk factors for the development of a recalcitrant nonunion were plate fixation of the acute humeral fracture, a history of deep infection, and ≥2 prior procedures. Plate fixation with bone graft augmentation remains a successful method for the treatment of humeral shaft nonunions. Neither plate type nor graft type reduced the risk of a recalcitrant nonunion. Factors that predicted a recalcitrant nonunion were operative fixation of the acute fracture with a plate, a history of deep infection, and ≥2 surgical procedures. Therapeutic Level IV. See Instructions for Authors for a complete description of levels of evidence.