Topologically trivial insulators come in two kinds: atomic, where the Wannier charge centers (WCCs) are localized on the atoms, and obstructed atomic, where the WCCs are located away from the atoms. The latter, which can exhibit interesting surface states and possibly have much larger band gaps than the topological insulators, have so far not been classified in three-dimensional (3D) crystalline materials. In this paper, we developed the 3D real space invariants (RSIs) for the 1,651 Shubnikov space groups with the spin-orbit coupling and provide the full classification of 3D obstructed atomic insulators (OAIs) by the RSIs. We then apply the theory to the entire database of materials on the Topological Quantum Chemistry website and Topological Magnetic Materials website, obtaining all the paramagnetic and magnetic OAIs so far existing in nature. We find that, out of the 34,013 paramagnetic and 296 magnetic topologically trivial insulators, there are 3,383 paramagnetic and 30 magnetic OAIs. All of them present a filling anomaly under certain open-boundary conditions and exhibit obstructed surface states (OSSs). We then derive the Miller indices of the cleavage planes which show the OSSs for all the OAIs and pick some of the best examples with large band gap to showcase their OSSs. We further refine the atomic insulator concept to obtain the orbital-selected OAIs (OOAIs), where the WCC of the system is located at a Wyckoff position occupied by an atom but forms a symmetric representation (orbital) that does not belong to the outer-shell electrons of the given atom. In such a way, we obtain a further 121 OOAIs. Furthermore, we analyze the catalytic properties of one of the OAIs in a "proof of principle" experiment. The surface of a high-performance heterogeneous catalyst is characterized by high stability, good electrical conductivity, and high charge carrier density near the Fermi level. These are also characteristic properties of the OSSs. By using a high-quality, single crystal of 2H-MoS 2 , that has well-defined surfaces, as a hydrogen evolution catalyst, we directly proved that the catalytic activities arise from the surfaces with OSSs, which is consistent with previous results based on electronic-structure calculations. Additional potential applications of the 3D RSIs and OAIs in, for example, electrochemistry, asymmetric catalysis, superconductivity and Josephson diode will be discussed.