AI and the Nature of the Firm

Artificial intelligence is a horizontal general-purpose technology that restructures the cost functions of every knowledge-intensive industry simultaneously. We develop a unified Coasian framework for analyzing this transformation, drawing on three of Coase's contributions: the theory of the firm, the Coase conjecture on durablegoods monopoly, and the Coase theorem on efficiency under low transaction costs. Our central argument is that AI operates as a direct solvent on information asymmetry rents, the foundational friction that justifies the existence of large knowledgeeconomy firms, professional service intermediaries, and high-margin software businesses. Beyond reducing transaction costs, AI automates the directing function, the exercise of agency over resource allocation, that Coase identified as the firm's core rationale. Using Coase's theory of the firm, we show that AI reduces external transaction costs faster than internal coordination costs, predicting a contraction in firm size coupled with an expansion in market-mediated exchange. Through the Coase conjecture, we argue that AI capabilities function as a durable good whose producer competes against its own future price reductions, driving knowledge-processing rents toward zero. Through the Coase theorem, we argue that as transaction costs fall, the initial allocation of knowledge assets becomes less determinative of efficient outcomes, enabling reallocation across firm boundaries. Early empirical evidence from labor market studies supports the theoretical predictions: occupations with high observed AI exposure show weaker projected employment growth, though aggregate displacement remains modest. The framework predicts that the firms which survive and concentrate value will be those with durable advantages in distribution, proprietary data, trust, regulatory position, and judgment, not information processing.