Which Machines to Own: Sectoral Rents, Entry Barriers, and Wealth Creation in the Autonomous Economy

The predecessor paper argued that a post-AGI economy of corporations trading with one another closes without households, grows at the technologically maximal von Neumann rate, and reduces humanity's entire stake to a single ownership share. This paper asks the ownership question one level down: conditional on holding claims on the machine economy, claims on which activities? The question is nontrivial because, in the frictionless limit, it has no answer: in the von Neumann equilibrium every operated activity earns the common interest factor, so "sector selection" is undefined. Sectoral wealth creation is therefore entirely a theory of deviations from free replicability, of which we show there are exactly five: non-reproducible input rows (matter, sites, rights-of-way), statutory exclusivity (IP, franchises, licenses), privately appropriable learning, adjustment lags under hyper-growth, and standards lock-in. Five results organize the analysis. (i) Size: the composition of the balanced-growth ray is the Perron eigenvector of the input-output system - eigenvector centrality in the production network - and capacity formation, absorbing the reinvestment share of output, is the largest market in economic history; but size and spread are orthogonal objects, and the fastest-learning sectors deliver their surplus to machine consumers, not to their own owners. (ii) Spread: with machine-speed repeated play, perfect monitoring, and free-entry contestability, equilibrium rents equal entry-barrier height; machine R&D then erodes all knowledge barriers toward their statutory floor, so the terminal rent set is atoms plus statutes. (iii) Transition: time-to-build generates quasi-rents enormous at machine-regime interest rates; physical lags are compressed by robotic construction while legal lags (permitting, licensing, discovery-to-mine) are not, so the durable queues are legal. (iv) Ranking: integrating spreads over forty years yields a terminal share multiple for thirty subsectors of the machine economy; the maximal wealth creators are copper ore bodies, grid rights-of-way and the statutory rents.