The high malignancy of glioblastoma has been recently attributed to the presence, within the tumor, of glioblastoma stem cells (GSC) poorly responsive to chemo- and radiotherapy. Here, the potential employment of metformin and arsenic trioxide (ATO) in glioblastoma therapy is discussed focusing on their effects on GSC. Metformin exerts anticancer effects by primarily blocking the pivotal LKB1/AMPK/mTOR/S6K1 pathway-dependent cell growth, induces selective lethal effects on GSC by impairing the GSC-initiating spherogenesis and inhibits the proliferation of CD133+ cells, while having a low or null effect on differentiated glioblastoma cells and normal human stem cells. Metformin and ATO induce autophagy and apoptosis in glioma cells by inhibiting and stimulating the PI3K/Akt and the mitogen-activated protein kinase pathways, respectively. Both drugs promote differentiation of GSC into non-tumorigenic cells. In this regard, metformin acts via activation of the AMPK-FOXO3 axis, whereas ATO blocks the interleukin 6-induced promotion of STAT3 phosphorylation. Blood-brain barrier, easily crossed by metformin but not by ATO, undergoes important glioblastoma-induced alterations that increase its permeability, thus allowing ATO to distribute more into the glioblastoma bulk than in the normal brain parenchyma. A prompt clinical assessment of metformin and ATO in glioblastoma patients would represent a valid attempt to improve their survival.