Friction and wear are common phenomena that impact all the applications where moving components are in contact, from micro-electromechanical systems to wind turbines, and result in massive economic and environmental costs. By advancing tribological materials impressive energy savings, and consequent reduction of CO2 emissions, can be obtained. However, optimizing lubricant materials is extremely challenging because their performances are ruled by molecular-level processes that occur at the buried interface, extremely difficult to monitor by experiments. Simulations can play a key role here, in particular those based on quantum mechanics, which is essential to accurately describe reactions in conditions of enhanced reactivity as those imposed by the mechanical stresses applied. Our group adopted a pioneering approach by applying ab initio MD for the first time to study tribochemistry. Since then we have established several collaborations with industries and experimental groups to discover improved materials to reduce friction by combining in silico and real experiments.