The experimentally observed planar hypercoordinate carbon species were detected in gas phase experiments and characterized by photoelectron spectroscopy. According to the Boltzmann distribution law, the thermodynamically favorable isomers, especially global minima, were relatively easier to detect than other isomers in such an experimental process. Here, we reported a thermodynamically unfavorable case, i.e., D-3h CN3Mg3+ (1a), which we think is experimentally viable because all isomers that are energetically lower than 1a show bimolecular assembly type structures consisting of an N-2 unit and various types of CNMg3+ units. The natural bond orbital (NBO) analysis suggests that the bonding between N-2 and CNMg3+ is rather weak, and we think it is very hard to retain their basic structures when kinetic factors are considered. Consistently, the four lowest isomers in the second group show dissociation (to free N-2 molecule and CNMg3+ cations) during Born-Oppenheimer molecular dynamic (BOMD) simulations. In contrast, the structure of 1a can be maintained under temperatures up to 2000 K during the BOMD simulation, and ring-opening reaction studies suggest the barrier to be very high, 46.75 kcal/mol. We think the excellent kinetic stability of 1a will compensate for its thermodynamic instability and it will own its existence in the gas phase synthesis. Although many isomers in the second group are energetically more favorable than 1a, they will be dissociated by the kinetic process. In the magnetic field, the positively charged CNMg3+ units will be separated quickly from N-2 molecules in the general gas phase synthesis, and they are therefore undetectable.