Isolated manganese atoms have half-filled 3d shells and a magnetic moment of 5 |iB. If this moment could be realized in permanent magnets, it would revolutionize technology. The present-day room-temperature record is about 2.43 |iB in Fe6 5Co3 5, and the maximum energy product of permanent magnets is quadratic in the magnetization. Furthermore, the crystal structure of Fe6 5Co3 5 is cubic and not suitable for permanent magnets, and it may be possible to create noncubic Mn magnets with substantial magnetocrystalline anisotropy. Known Mn-based permanent magnets, such as MnAl, MnBi, and Mn2Ga, exhibit modest energy products. The main reason is the low net magnetic moment, of the order of 0.5 |iB per atom, which means that 99% of the energy product are wasted. The focus of our presentation is the theoretical explanation of the low moment of Mn-based permanent magnets and the search for crystal structures with improved net magnetization. Analytical quantum mechanics and density functional (DFT) calculations  will be used to tackle the problem. Since the magnetization is equal to the moment per volume, the Mn atoms need to be compacted into solids, and the resulting interatomic hybridization tends to reduce the atomic moments. However, moments of about 3 |iB are rather easy to realize in solids, and up to about 3.7 |iB per atoms are not unrealistic even in metals . Another factor is the dilution ofthe magnetization due to the presence of nonmagnetic atoms, e.g., Mn moments approaching 5 |iB can be realized in oxides, but the large volume ofthe O2 - ions severely limits the magnetization.