Generative AI platforms as institutional catalysts of digital entrepreneurship : enablement, dependence & power dynamics

Abstract

Altres ajuts: acords transformatius de la UAB

Altres ajuts: ICREA Academia programme

This study theorizes how recent generative AI (Gen AI) platforms operate as institutional catalysts of platform-dependent entrepreneurship (PDE). Integrating institutional theory, the external enablement framework, and innovation platform theory, we propose an integrative framework for explaining the emergence of PDE under reliance on non-substitutable, platform-governed capabilities. Using an abductive mixed-method case study of OpenAI's ecosystem (2020-2025), we trace how governance, boundary resources, and institutional signals shape entrepreneurial feasibility, scaling, and vulnerability. Our analysis identifies four catalytic institutional mechanisms-Infrastructure Provision, Capability Scaffolding, Market Legitimization, and Ecosystem Orchestration-that enable venture creation while simultaneously generating dependence. Temporal analysis reveals an enablement-dependence paradox: platforms accelerate entry by democratizing frontier capabilities, yet accumulate dependencies that expose ventures to governance shocks. Empirically, we show how enthusiasm gave way to crisis during OpenAI's GPT-5 release, illustrating governance overreach, trust erosion cascades, and choice removal as control. We conclude with theoretical and practical implications.