Digital Twins are transforming industrial systems by enabling real-time monitoring, predictive analytics, and intelligent control through the integration of physical assets with their digital counterparts. As industrial applications become more complex and data-rich, scalable computational methods—often augmented by artificial intelligence and machine learning—are extremely important for the realization of digital twins with guaranteed performance, adaptability, and reliability. This minisymposium focuses on recent advances in scalable computational methods that support the development and deployment of Digital Twins in industrial settings. We invite contributions that address the algorithmic, methodological, and implementation challenges associated with high-fidelity, real-time, and large-scale digital twin applications. Topics of interest include, but are not limited to: • Model order reduction and surrogate modelling for real-time simulation in industrial-scale applications • Uncertainty quantification and robust model calibration for reliability and trustworthiness • AI- and ML-based data-driven modelling techniques for enhanced autonomy and adaptability We particularly encourage submissions that demonstrate the application of these or similar methods to real-world industrial systems, highlighting the synergy between computational innovation and practical impact.