Results of first‐principles, density‐functional LMTO‐ASA calculations on the two title compounds and of unrestricted Hartree‐Fock (UHF) calculations on VF2 are reported. The calculations on the crystalline materials were performed with special emphasis on optimising all (three) structural parameters, on the valence density of states and the valence electron density, and on the possible existence of spin polarisation. In addition, full‐potential LMTO calculations on isolated monomers were carried through in order to examine the nearest‐neighbour (TiO and VF) chemical bonds in detail and to understand the consequences of various approximations in describing the potential for the crystal. The LMTO‐ASA calculations reproduce the correct lattice constants with, however, some overestimate for VF2. Also the nearest‐neighbour TiO and VF bond lengths are reproduced correctly by the LMTO calculations, whereas the c/a ratios are significantly overestimated. By analysing the charge densities of the isolated monomers we conclude that the overestimates in the c/a ratios are due to the ASA (Atomic‐Sphere Approximation), which does not describe the quadrupole‐quadrupole interactions of the compounds sufficiently accurately. The Hartree‐Fock calculations yield lattice constants for VF2 very close to experiment. Also in agreement with experiment, TiO2 is found to be non‐magnetic and insulating/semiconducting. In contrast, VF2 is found to be metallic within the LMTO‐ASA method, although the density‐functional calculations indicate that magnetic effects will change the compound into an insulator as it is according to experiment. On the other hand, in agreement with experimental evidence, UHF yields a non‐metallic structure. It further shows the stability of a number of patterns of magnetic ordering.