Glioblastoma multiforme (GBM) is the most common and deadliest primary brain tumor. Immunotherapeutic approaches using chimeric antigen receptor (CAR) T cells have shown limited efficacy against GBM due to heterogeneous target antigen expression. We hypothesize that human cytomegalovirus (HCMV) can serve as a therapeutic target for GBM. HCMV can be detected in up to 90% of GBM tumor samples but not the surrounding normal brain tissue. The role of HCMV as a tumor-promoting virus is poorly understood but its presence in the tumor presents a novel approach to developing a therapy for GBM by re-directing T cells to target HCMV. We detected the presence of HCMV in GBM tumor samples via IHC and confirmed expression of HCMV gene UL55/glycoprotein B (gB) in 45% of primary GBM tumors. A CAR was generated and optimized to recognize the HCMV surface antigen gB. In vitro testing of the anti-gB CAR revealed activity against the U87 glioma cell line stably transduced to express gB and CMV-infected human foreskin fibroblasts (HFF) cells. In vivo, gB CARs were able to treat established GBM tumors in a xenograft mouse model. In vitro co-cultures of gB CAR T cells against the human GBM explant, D270, demonstrated tumor recognition and anti-tumor function against primary GBM. gB CAR T cells were able to control D270 tumor growth in vivo despite undetectable levels of antigen expression. Mice displaying stable disease showed improved persistence of engrafted human T cells and tumor infiltration. These results suggest that CAR T cells may be effective in recognizing extremely low abundance antigens, and taken together, the results of this study show the feasibility of using gB CAR T cells as a platform to target HCMV in GBM tumors to treat patients with GBM. Ultimately, the goal of this study is translate these findings into clinical trials.