Illustrating by compelling case studies, this study discusses the challenges to the design of resilient power grid with inverter-based renewable energy resources posed by curses of dimensionality and heterogeneity in the complex multi-scale interactive dynamics by utilizing the network-based data-driven approaches, and further explains the essentiality of having hybrid data analytics with multi-scale stochastic autocovariance analysis to overcome those newly recognized curse of varietal complexity in next-generation power grids. It underscores the necessity for resilient functions beyond traditional reliability, explores essential resiliency aspects of power grids integrated with renewable resources, emphasizing proactive mechanisms to ensure stability during disruptions. By integrating numerical methods and differential embedding techniques, the framework enables precise modeling of energy flow dynamics across temporal and spatial scales, crucial for accurate transient analysis. Simulation case studies further illustrate its capability in distinguishing nonlinear dynamics induced by IBRs from traditional synchronous generators, showcasing robustness and accuracy even under varying penetration levels of IBRs. This innovative approach not only enhances modeling accuracy and prediction capabilities but also provides insights essential for managing and controlling the complex dynamics of modern power grids.