Neuromorphic computing represents a revolutionary approach to developing analog non-volatile memory (NVM) devices. This technology relies on electron-correlated materials capable of replicating the functioning of synapses and neurons in the brain, making it ideal for energy-efficient parallel computing applications. Among the myriad of innovative materials suitable for neuromorphic computing, such as conductive polymers, organic semiconductors, and nanowires, among others, electron-correlated transition metal oxides stand out for their ability to undergo metal-insulator transitions either by tuning their oxygen content or the stoichiometry of an intercalating metal ion. In this talk we discuss from a first principles perspective the pivotal role that vanadium oxide has the potential to play in the development neuromorphic computing devices. First, we examine the use of vanadium oxide as a Mott insulator capable of undergoing a stable phase separation process that is suitable for non-volatile electrochemically activated random access memory devices (ECRAM). A stable phase coexistence across multiple segregated vanadium oxides states provides the device the ability express non-volatile memory behavior after precise tunning of oxygen vacancies into VO2. The correlation between these oxides’ formation energies and their oxygen content provides clues on the stable formation of coexisting crystalline phases that provide tunable resistivity based on the relative fraction and distribution of each phase with unusual retention, with an estimated 1% loss over 14 years in ambient conditions. Second, we address the structural and electrical properties of β’-CuxV2O5 based on temperature and Cu content. Ab-initio molecular dynamics calculations provide a close-up view of the correlation between Cu content and mobility and its ability to induce structural distortions to the intercalating β’-CuxV2O5 host that spike close to the insulator-metal transition (IMT).