Rapid advances in artificial intelligence (AI) and machine learning are transforming materials development through data-driven prediction, materials exploration, and accelerated simulation. However, the properties of real materials are affected not only by their chemical composition but also by hierarchical structures spanning atomic, molecular, mesoscopic, and macroscopic scales, together with processing history. Physics-based simulation therefore continues to play an important role in materials design. This presentation reviews research on multiscale modeling accumulated through OCTA/J-OCTA and discusses its recent integration with AI, including structure recognition, surrogate modeling, and optimization. We also introduce MI-Suite, which aims to integrate experimental and simulation data for analysis and materials exploration. Through these examples, we discuss how multiscale simulation can evolve and complement AI in future materials development.