Cognitive Capital and AI: Who Benefits, How Fast, and How Broadly

Economic growth is often explained by physical and human capital, yet the biggest gains from AI arrive only when organizations build the cognitive systems that let new tools actually work. We call this Cognitive Capital : the data architecture, domain knowledge, interoperable workflows, governance, and institutional learning that turn AI and information into reliable decisions and actions. We develop a task-based framework showing how Cognitive Capital reshapes which tasks expand, how labor is reallocated, and what gets traded across borders. The Three T's-Task Augmentation, Tradability, and Technology readiness-jointly determine whether economies move into more cognitive, AI-complementary work and experience "virtual migration" of services. We also open the black box of the information environment. Personalized feeds can improve private learning while eroding shared understanding, raising polarization. When polarization is high, investments in organizational capability convert less effectively and depreciate faster, creating "fragmentation traps" in which even large programs fail to lift productivity. Conversely, shared standards, data governance, and pro-openness norms amplify Cognitive Capital and speed diffusion. The framework delivers testable predictions for adoption speed, wage dispersion, export composition, and the emergence of multiple development paths. The central claim is simple: as with electrification and factory systems, it is not AI alone but the cognitive systems built around it that determine who benefits, how fast, and how broadly.