High-efficiency near-infrared (NIR) emitting materials play a crucial role in biomedicine, agriculture, spectroscopy, etc., and the stannates emerge due to their broadband and millisecond-lifetime NIR luminescence. However, their origin has confused the community for 20 years. In this work, we combine first-principles calculations with experiments to explicitly unveil that the unique NIR emission originates from the electronic transitions of trace impurities rather than the previously proposed Sn 2+ -related self-trapping excitons (STEs) transition in the Sn 4+based compounds, such as AESnO 3 (AE = Ba, Sr, and Ca) perovskites, Ca 2 RESn 2 M 3 O 12 (RE = Lu, Y, and Gd; M = Al; RE = La; and M = Ga) and Na 2 CaSn 2 Ge 3 O 12 garnets, Mg 2 SnO 4 inverse spinel, and La 2 Sn 2 O 7 pyrochlore. First-principles calculations provide detailed insights, ruling out the contribution of Sn-related activators to the NIR emission due to their high formation energy and strong electron−phonon coupling but confirming the contribution of Fe 3+ or Cr 3+ centers by analyses of site occupancies and 3d−3d optical transitions, which effectively explain the experimental emissions and their shift trends in various hosts, as well as the characteristics of luminescence decays. Furthermore, Fe 3+ -activated BaSnO 3 perovskite exhibits decent NIR radioluminescence (∼905 nm) and thus demonstrates it has potential as a new type of NIR scintillator. The fresh physical picture in our work thoroughly resolves the origin of stannates' NIR luminescence and extends the radioluminescence to the broadband NIR wavelengths with millisecond-scale lifetimes for time-resolved bioimaging and biosensing.