We investigate the capability of the strongest earthquakes to modify significantly the seismicity in a wide spatiotemporal window. In particular, we show that the strongest earthquakes of last century were probably able to influence the seismicity at large spatiotemporal distances, extending their reach over thousands of kilometers and decades later. We report statistically significant differences in worldwide seismicity before and after the occurrence of the strongest earthquakes of the last century, whose perturbation is modeled by means of coseismic and postseismic stress variations. This long-term coupling has produced time variations in worldwide seismic activity that appear related to the physical coupling between the focal mechanism of source earthquakes and the tectonic setting of each zone. These results could provide new important insights on seismic hazard assessment because they raise doubts on the validity of two paradigms; that is, the steadiness of the mainshock rate and the isolation of a seismic region from the surrounding areas. Finally, in addition to this backward analysis, we also provide a formal forward test by forecasting the effects on global seismicity of the recent Sumatra–Andaman earthquakes; this is maybe a unique chance to test the long-term hypothesis with an independent dataset that avoids, by definition, any kind of (often unconscious) optimization of the results that is an unavoidable possibility in backward analyses.