<p>Since the mid-1990s computing revolution, the U.S. economy has exhibited several striking regularities. Labor productivity growth has slowed below its historical trend following a decade-long boom, while the labor share has declined. Computing technologies are characterized by rapid obsolescence, and U.S. data suggest that aggregate capital obsolescence has accelerated since their widespread adoption. I study the macroeconomic implications of this acceleration using an endogenous growth model in which the economy transitions to a new equilibrium with a higher obsolescence rate. As capital designs become obsolete sooner, the value of future payoffs to capital innovation initially falls. The economy responds with a temporary productivity boom driven by faster skill accumulation, which partially restores innovation profitability. Over time, however, efficient capital per efficient labor declines. Under factor complementarity, this scarcity depresses the labor share by raising flow profits accruing to capital designs relative to skills. Resources are redirected from skill creation toward short-lived capital innovation, ultimately slowing long-run growth. The calibrated model produces productivity and labor-share dynamics broadly consistent with the data, and the paper advances the obsolescence channel as one plausible account of these trends.</p>