Abstract Background Human immunodeficiency virus type 1 (HIV-1) remains a persistent global health challenge. Therefore, a continuous exploration of novel therapeutic strategies is essential. A comprehensive understanding of how HIV-1 utilizes the cellular metabolism machinery for replication can provide insights into new therapeutic approaches. Methods In this study, we performed a flux balance analysis using a genome-scale metabolic model (GEM) integrated with an HIV-1 viral biomass objective function to identify potential targets for anti–HIV-1 interventions. We generated a GEM by integrating an HIV-1 production reaction into CD4+ T cells and optimized for both host and virus optimal states as objective functions to depict metabolic profiles of cells in the status for optimal host biomass maintenance or for optimal HIV-1 virion production. Differential analysis was used to predict biochemical reactions altered optimal for HIV-1 production. In addition, we conducted in silico simulations involving gene and reaction knock-outs to identify potential anti–HIV-1 targets, which were subsequently validated by human phytohemagglutinin (PHA) blasts infected with HIV-1. Results Differential analysis identified several altered biochemical reactions, including increased lysine uptake and oxidative phosphorylation (OXPHOS) activities in the virus optima compared with the host optima. In silico gene and reaction knock-out simulations revealed de novo pyrimidine synthesis, and OXPHOS could serve as potential anti–HIV-1 metabolic targets. In vitro assay confirmed that targeting OXPHOS using metformin could suppress the replication of HIV-1 by 56.6% (385.4 ± 67.5 pg/mL in the metformin-treated group vs. 888.4 ± 32.3 pg/mL in the control group, P < 0.001). Conclusion Our integrated host-virus genome-scale metabolic study provides insights on potential targets (OXPHOS) for anti-HIV therapies.