Replication Data for: Decoding National AI Regulatory Capacity: A Mixed-Methods Study at a Global Scale

The proliferation of artificial intelligence (AI) brings significant risks that require robust national regulatory capacity for mitigation. However, the mechanisms through which this capacity forms remain undertheorized. Addressing this gap, our study employs a sequential mixed-methods approach, integrating QCA and NCA applied to 135 countries with longitudinal process-tracing of four representative cases. We develop a novel endowment-preference-incentive (EPI) framework to assess AI regulatory capacity globally, identifying four equifinal paths along a spectrum from reaction to proactivity: interventionist reaction, (centralized or defensive) harmonist responsiveness, and entrepreneurial proactivity. These paths generate distinct regulatory regimes, characterized by varying approaches to regulatory foundations, standard-setting, and penalty imposition. This study contributes to AI governance, regulatory theory, and comparative governance by elucidating the causal mechanisms behind national AI regulatory capacity formation, presenting a robust typology for comparative research, and reconceptualizing regulation as a dual-purpose instrument that both enables innovation and safeguards public interests and technological sovereignty.