The cancer stem cell (CSC) hypothesis suggests that tumours are maintained by a subpopulation of cells with stem cell properties. Although the existence of CSCs was initially described in human leukaemia, less evidence exists for CSCs in solid tumours. Recently, a CD133+ cell subpopulation was isolated from human brain tumoursexhibiting stem cell properties in vitro as well as the capacity to initiate tumours in vivo. In the present work, we try to summarize the data showing that some elements of the Phosphoinositide 3-kinase Class I (PI3K)/ Thymoma viral oncogene protein kinase (Akt) pathway, such the activity of PI3K Class I or Akt2, are necessary to maintain the CSC-like phenotype as well as survival of CSCs (also denoted as tumour-initiating cells (TICs)). Our data and other laboratory data permit a working hypothesis in which each Akt isoform plays an important and specific role in CSC/TIC growth, self-renewal, maintaining survival, and epithelial-mesenchymal transition (EMT) phenotype, not only in breast cancer, but also in glioma. We suggest that a more complete understanding is needed of the possible roles of isoforms in human tumours (iso-signalling determination). Thus, a comprehensive analysis of how hierarchical signalling is assembled during oncogenesis, how cancer landmarks are interconnected to favour CSC and tumour growth, and how some protein isoforms play a specific role in CSCs to ensure that survival and proliferation must be done in order to propose/generate new therapeutic approaches (alone or in combination with existing ones) to use against cancer.