Cumulative r − g, 2010–2030
+342 pp
Sum of annual gaps where r exceeds g. The wealth-concentration premium for AI capital, against the broad economy.
Innovation Economics · Twin time series
r > g, applied to AI capital.
Piketty's central empirical claim: when the rate of return on capital (r) exceeds the rate of growth (g), wealth concentrates. AI is a near-perfect r > g technology — fixed costs are enormous, marginal cost of inference is near zero, returns to scale are substantial, the compounding advantage of data and infrastructure access is large. Plotted: real global GDP growth against the return on a frontier-AI capital basket. The shaded gap is the concentration mechanism.
Years where r > g
20 / 21
In this 21-year window, AI capital out-earned broad GDP growth in nearly every year.
The thesis
Wealth concentration is the inequality dimension that the growth-rate-focused models systematically miss. Piketty's data show that even in periods of strong productivity growth, capital concentration can rise if r continues to exceed g. The institutional response is well known — progressive wealth and inheritance taxation, antitrust, broad ownership models — but it requires deliberate choices that most jurisdictions, including South Africa, have not yet made.