The Infrastructure of Power: Finance and Artificial Intelligence

The last four decades of global capitalism have been defined by a single structural logic, that of power exercised through the control of capital allocation. Financial infrastructure (investment banks, asset managers, sovereign wealth funds) has dominated this logic, determining which projects, technologies, and geographies receive resources and which do not. This paper argues that a structural transition is underway. Artificial intelligence, initially a tool deployed within financial infrastructure, is progressively redefining the horizon inside which that infrastructure reconfigures itself. Finance does not disappear, nor is it absorbed by AI; what changes is the field of possibility within which financial actors think, formulate problems, and identify solutions. Drawing on empirical evidence from the period 2022–2026, including the financialization of compute as an asset class, the operational penetration of large language models across high finance, the concentration of venture capital toward foundational model layers, the migration of financial risk categories into computational logic, and the replication of pre-2008 systemic risk geometry in cognitive infrastructure, the paper identifies the structural mechanism of this transition and its implications for the distribution of global power. The analysis is conducted from the standpoint of someone moving between both domains, a position that proves not merely biographical but epistemically necessary, since the transition is visible only from within the intersection it itself produces.

Keywords: Financial infrastructure, cognitive infrastructure, artificial intelligence, capital allocation, foundational models, systemic risk, power, surveillance capitalism, compute, venture capital concentration, fintech.