Empirical studies demonstrate that interbank networks are modular and scale-free. This study applies the contagion framework developed by Nier et al. to assess how realistic topologies shape systemic risk. While Nier et al. employed Erdős–Rényi (ER) networks with homogeneous connectivity, we generate scale-free networks via preferential attachment and introduce modularity through the probabilistic rewiring method of Nguyen et al. The comparison reveals fundamental differences: modular scale-free networks exhibit gradual stabilization driven by the shock-absorbing and localizing roles of hubs, with persistent contagion explained by clustered neighborhoods and fragile peripheral nodes. In contrast, ER networks display synchronized stabilization around critical thresholds. Targeted shocks expose the dual role of hubs, whose failure amplifies contagion across modules and reflects the “robust-yet-fragile” nature of scale-free systems. These findings underscore the importance of network topology in evaluating systemic risk.