The ambition is to predict the most promising catalyst for CO2, CH4, and NxO conversion either on homogeneous or heterogeneous phase. The specific research objectives are to predict the fixation and functionalization of greenhouse gases applying predictive catalysis; and to explore new routes for H2 production without a carbon footprint. In particular, the CO2 cycloaddition to epoxides to give rise to cyclic carbonates has attracted much attention, trying to find out the catalysts and cocatalysts that enhance the performance of this reaction. In addition, other strategies will be tested to fixate and/or functionalize CO2 by means of Frustrated Lewis Pairs, or N2O by N-heterocyclic carbenes or both gases by diborylstannylene complexes. It is a two-way project. First, because it aims to be a meeting point from homogeneous catalysis to heterogeneous catalysis but also vice versa. Second, at the same time, it will also be multidisciplinary, as it will include experiments and then calculations, in this order, but here above all the order will be the reverse, making intensive use of predictive catalysis. This concept may seem distant and almost unattainable as machine learning is still due to the lack of data, but we must forget about this, we need to change the paradigm from now on and make calculations replace experimental scans, and that guide them to greater efficiency, while reducing the amount to save resources. But we must not forget that supercomputers consume energy, and this must be said in all honesty, so the green energy itself must sustain them in the near future.