Nonunion is defined as the permanent failure of a fractured bone to heal, often necessitating surgical intervention. Atrophic nonunions are a subtype that are particularly difficult to treat. Animal models of atrophic nonunion are available; however, these require surgical or radiation induced trauma to disrupt periosteal healing. While these approaches can result in nonunion, such invasive methods are not representative of many clinical nonunions where osseous regeneration has been arrested by a “failure of biology”. We hypothesized that arresting osteoblast cell proliferation after fracture would lead to atrophic nonunion in mice. Using mice that express a thymidine kinase (tk) ‘suicide gene’ driven by the 3.6Col1a1 promoter (Col1-tk), proliferating osteoblast lineage cells can be ablated upon exposure to the nucleoside analog ganciclovir (GCV). Wild-type (WT; control) and Col1-tk littermates were subjected to a full femur fracture and intramedullary fixation at 12 weeks old. Post injury, mice were dosed with GCV twice daily for 2 or 4 weeks. Histologically, we confirmed abundant tk+ expression in fracture callus, and diminished periosteal cell proliferation in Col1-tk mice at 3 weeks post fracture. Moreover, Col1-tk mice had less osteoclast activity, mineralized callus, and vasculature at the fracture site compared to WT mice. Additional mice were monitored for 12 weeks with in vivo radiographs and microCT scans, which revealed delayed bone bridging and reduced callus size in Col1-tk mice. Following sacrifice, ex vivo microCT and histology demonstrated failed union with residual bone fragments and fibrous tissue in Col1-tk mice. Biomechanical testing demonstrated an inability to recover torsional strength in Col1-tk mice compared to WT. Our data indicates that suppression of proliferating osteoblast-lineage cells for either 2 or 4 weeks after fracture blunts the formation and remodeling of a mineralized callus leading to a functional nonunion. We propose this as a new murine model of atrophic nonunion.