Additive manufacturing (AM), and in particular laser-based processes such as Direct Energy Deposition (DED), require accurate simulations to predict and prevent defects. Thermal evolution is the main physical phenomenon that drives mechanical deformation and metallurgical changes. These processes are inherently multi-scale: highly localized and transient thermal gradients arise near the moving heat source, while the global thermal response governs residual stresses and distortions at the component scale. Capturing both effects efficiently remains a major computational challenge. The AM industry currently faces significant limitations when it tries to simulate DED processes numerically. Most of the available simulation tools are general-purpose finite element software,that is not specifically designed for AM technology. Consequently, simulations are often restricted to simplified geometries, low accuracy approaches, or small components. Part-scale simulations remain computationally prohibitive within this framework. State of the art simulations [1] rely on finite element activation algorithms that use a mesh of the entire component. Using this approach, increasing the accuracy over the critical region around the heat source requires the entire mesh to be refined. As an alternative, we propose an Arlequin-based method [2], that uses a small fine moving mesh along with a large coarser static mesh to perform the simulations. Each mesh is designed to capture one of the two aforementioned main thermal effects previously mentioned. The talk aims to evaluate the scalability and computational efficiency of the method. Numerical results will be presented and compared with those from classical finite element implementations. The results will demonstrate the potential of the proposed approach in enabling accurate and cost-effective thermal simulations at the component level for DED AM processes.