Hotspot fixity and no major plate reorganization in the past ∼40 m.y. are two propositions important for studying absolute plate motion and mantle dynamics. Reexamining the hotspot data with the NUVEL-1A relative plate motion model, we find that these two propositions cannot be simultaneously tenable: either hotspots have been moving relative to each other, or a major reorganization of plate motion occurred in the past ∼40 m.y. Statistical compatibility tests show that hotspot rate data are incompatible with trend data, implying that hotspots have moved. Furthermore, hotspot rates are consistently lower than those predicted by best-fitting absolute plate motion models. This may be explained by hotspots moving systematically opposite to the plate motion. If so, the moving hotspots still provide a useful reference frame for defining absolute plate motion. However, all hotspot trend data, with ages as old as 40 Ma, are statistically compatible, indicating no major reorganization of plate motion in the past ∼40 m.y. In this case, relative plate motion models based on young marine magnetic anomalies may be used to infer older (<40 Ma) plate kinematics, which lack direct constraints.