Julian Xu

Abstract

Model of Models: A Configurational Analysis of Business Model Archetypes for Foundation Models using Platform Economics and Qualitative Comparative Analysis The rapid ascent of Generative AI presents a structural puzzle for platform economists: why does the Foundation Model (FM) market exhibit persistent strategic diversity despite high fixed costs that typically drive monopolistic convergence? This research addresses the question: What are the distinct, causally sufficient configurations of strategic conditions that produce market viability in the FM market, and why do multiple configurations coexist rather than converging to a single form? Drawing upon the PEMAK framework (Platform Economics: Matching, Assembling, and Knowledge Management), this study conceptualizes FMs as digital platforms performing specialized economic functions. Using Qualitative Comparative Analysis (QCA) on a global population of approximately 30 major model families, the research identifies a set of distinct, viable business model archetypes. By analyzing strategic dimensions—including market power, ecosystem integration, multimodality, pricing, governance, and openness—the study uncovers a landscape of equifinality where divergent strategies achieve equivalent market viability. Key findings include the identification of stable strategic "recipes" and a significant "structural absence"—configurations that are theoretically possible but empirically unviable due to the capital-intensive nature of the FM layer. The research contributes to scholarly literature by formalizing the functional identity of AI as a platform and providing a strategic heuristic for practitionners and policymakers to navigate the AI-as-a-platform economy. This study challenges the assumption of convergence in winner-take-all markets and offers a predictive map for the evolution of the AI industry.