Background One of the main challenges in cancer treatment is addressing the metabolic reprogramming of tumor cells, which require more energy and biomolecules than healthy cells. Cancer cells alter their metabolism by switching between glycolysis and oxidative phosphorylation (OXPHOS). These processes depend on transmembrane proteins that respond to the extracellular environment. Our research identified the transmembrane form of Chloride Intracellular Channel 1 (tmCLIC1) as a marker of malignancy and a potential therapeutic target. tmCLIC1 levels are increased in several solid tumors, supporting cancer growth and progression, whereas they are mostly absent in healthy cells. We confirmed that tmCLIC1 is the specific target of the antidiabetic drug metformin, an OXPHOS inhibitor in cancer cells. Methods tmCLIC1 is the primary target of metformin in glioblastoma-initiating cells, as shown by single-channel patch-clamp recordings and NMR experiments. Various patient-derived glioblastoma cells with different genetic backgrounds were used to demonstrate that CLIC1 CRISPR-Cas9 knockout and/or its point mutation at arginine 29 removes metformin’s antitumor effects. Functional assays were used to assess the effects on proliferation, mitochondrial metabolism, and tumor growth in vitro and in vivo, using zebrafish and murine xenograft models. Results Metformin inhibits the function of tmCLIC1 through direct and specific binding involving arginine 29 in the tmCLIC1 sequence. Additionally, during hypoglycemia, metformin promotes glioblastoma cell apoptosis by inhibiting the Cancerous Inhibitor of Protein Phosphatase 2 A (CIP2A) and activating the PP2A B56δ subunit. This leads to the dephosphorylation of Glycogen Synthase Kinase 3 Beta (GSK3β), resulting in the breakdown of the pro-survival protein MCL-1 and subsequent cell death. Inhibition of tmCLIC1 is crucial for this metformin-driven antineoplastic effect, mainly through regulating the PP2A-GSK3β-MCL-1 pathway under hypoglycemic conditions. The chronic presence of metformin within the tumors impairs in vivo growth at nanomolar concentrations. Conclusions The therapeutic role of metformin to treat brain tumors remains debated. Our findings show that drug delivery is essential, as in vivo, tumor growth decreases at concentrations below 10 nanomolar. We propose that sustained CNS metformin levels may improve tmCLIC1 inhibition, providing a basis for optimizing interactions with metformin or related compounds to enhance therapeutic efficacy.