The study by Yin et al. (2021) describes the impact of data from the Chinese Geostationary Interferometric Infrared Sounder (GIIRS) on the forecast of typhoon Maria (2018). The study deserves highlighting because it is a milestone on the path of geostationary hyperspectral sounders (hereafter GeoHSS): GIIRS is the first GeoHSS and realizes what had been proposed by others for quite some time. GIIRS is deployed on the FY-4 series of Chinese geostationary meteorological satellites (Yang et al., 2017). The FY-4 series is a new generation of Chinese geostationary meteorological satellites and the first one, FY-4A, was launched on December 11, 2016. GIIRS has novel capabilities for vertical temperature and moisture sounding of the atmosphere; it has spectral channels in two bands (LW: 700-1,130 cm −1 for temperature and MW: 1,650-2,250 cm −1 for humidity and lower tropospheric temperature). Yin et al. (2021) use 48 channels from the LW band. The spatial resolution is 16 km at sub-satellite point. Technical improvements are foreseen for the next FY-4 satellites (Yang et al., 2017). Readers interested in a comprehensive description of the evolution of infrared sounding are referred to Menzel et al. (2018). My commentary consists of three parts: (a) I highlight the new potential tapped with GIIRS, (b) I recall earlier plans for GeoHSS which did not materialize yet helped the realization of GIIRS, (c) I conclude by referring to the vision of the World Meteorological Organization (WMO, 2020) which states the operational need for a ring of GeoHSS instruments around the Earth, similarly to what has been realized for imaging instruments. 1. Yin et al. (2021) describe the assimilation of targeted observations with GIIRS of Typhoon Maria which made landfall over Fujian, China on July 11, 2018. Radiance observations of selected channels were used in the Global/Regional Assimilation and PrEdiction System (GRAPES)-Global Forecast System (GFS) 4D-Var system (Shen et al., 2020). For the first time, the beneficial impact of a GeoHSS on forecasting of a tropical storm is demonstrated. Results are improved track forecast and reduced intensity forecast errors. Furthermore, the assimilation of observations with different temporal resolutions (from 15 min to