Titanium nitride (TiN) is a widely used hard coating in surface engineering. This material is characterised by a number of advantageous properties, including elevated hardness, excellent wear resistance, and superior thermal stability [1]. However, despite its technological relevance, the brittle fracture behaviour of TiN at the nanoscale remains insufficiently characterised, particularly with respect to the reliability of numerical predictions of fracture initiation and propagation. That is why this study presents a comprehensive numerical investigation of fracture behaviour in nanocrystalline TiN, with an emphasis on quantifying fracture resistance and evaluating size effects in atomistic simulations. Such results can then be used as input to the large-scale fracture model, thereby creating the multiscale framework. The current work is based on large-scale molecular dynamics (MD) simulations [2] with the LAMMPS library, in which fracture processes under uniaxial tensile loading are examined using pre-cracked geometrical models. Interatomic interactions are described by a potential specifically selected for its capability to capture the directional bonding characteristics of TiN. To evaluate the influence of system dimensions on the computed fracture response, a broad spectrum of model sizes is investigated. Fracture toughness is finally determined using both force-based and energy-based methodologies. Furthermore, a numerical stability analysis at the atomic scale is conducted to identify and mitigate unrealistic stiffening effects and to prevent significant overestimation of the fracture resistance. The results underscore the critical importance of appropriate model size selection in atomistic fracture simulations and highlight the necessity of large-scale parallel computations for the reliable prediction of fracture behaviour in brittle ceramic coatings. The proposed computational framework establishes a foundation for subsequent studies of more complex coating–substrate systems, which are of key importance for advanced engineering applications and for numerical analyses based on the finite element method [3].