Evidence that the automation of routine tasks has contributed to the polarization of labor markets has been documented for many developed economies, but little is known about its incidence in developing economies. We propose a measure of the exposure to routinization—that is, the risk of the displacement of labor by information technology—and assemble several facts that link the exposure to routinization with the prospects of polarization. Drawing on exposures for about 85 countries since 1990, we establish that: (1) developing economies are significantly less exposed to routinization than their developed counterparts; (2) the initial exposure to routinization is a strong predictor of the long-run exposure; and (3) among countries with high initial exposures to routinization, polarization dynamics have been strong and subsequent exposures have fallen; while among those with low initial exposure, the globalization of trade and structural transformation have prevailed and routine exposures have risen. Although we find little evidence of polarization in developing countries thus far, with rapidly rising exposures to routinization, the risks of future labor market polarization have escalated with potentially significant consequences for productivity, growth and distribution.