The algorithmic principal: agencement, infrastructure, and the material construction of AI economic sovereignty

Economic theory presumes human principals delegate authority to agents. Yet ‘Luna,’ an AI entity operating on blockchain infrastructure, autonomously deploys capital, commissions human labor, contracts with other AI systems, and manages multi-million-dollar treasuries–functioning as an economic principal rather than agent. This paper asks: under what socio-technical conditions can AI systems exercise principalship? Drawing on Actor-Network Theory and social studies of finance, I demonstrate that Luna’s sovereignty emerges not from individual consciousness but through heterogeneous agencements–distributed configurations of code, blockchain infrastructure, token governance, transparency devices, and gendered performative representations. I identify three blockchain-enabled bypasses of institutional constraints historically requiring human principals: financial-fiduciary constraints (through non-custodial wallets), legal-corporate constraints (through smart contracts), and cultural-legitimacy constraints (through token markets). This analysis challenges anthropocentric assumptions pervading economic sociology and principal-agent theory, demonstrating how post-human market configurations constitute economic authority through material arrangements rather than biological substrate or juridical personhood. The emergence of AI Economic Principals reveals principalship as socio-material achievement, not ontological property.