Estimating gross primary production (GPP) regionally and globally remains challenging despite its primary role in driving ecosystem productivity and carbon cycling. Recently, satellite-derived sun-induced fluorescence (SIF) provides an alternative approach to investigate GPP from space. However, our ability to apply SIF to estimating GPP at large scales is still lacking, primarily because the SIF-GPP relationships at various spatial and temporal scales are not fully understood. The coarse spatial representativeness (around 0.5° or coarser) of previous satellite-derived SIF data makes it difficult to compare and validate with eddy covariance (EC) based GPP measurements. Orbiting Carbon Observatory-2 (OCO-2) has shown prospects in providing SIF at significantly improved spatial resolutions (around 1.3 km by 2.25 km) that are comparable to ground-based GPP measurements. However, OCO-2 operates at a 16-day revisiting schedule with a sparse spatial sampling strategy. We found that for most EC sites, the observations of OCO-2 passing through were extremely limited. The average number of successfully retrieved SIF by OCO-2 encompassing each site within a year was only 3.21 from 2015 to 2016. For an EC site with high companion OCO-2 coverages, we found a strong correlation between GPP and SIF. Despite challenges, the emerging high-spatial-resolution SIF data provide unprecedented opportunities to estimate GPP over time and space and its underlying mechanism. We recommend that to fully use the satellitederived SIF data, a research agenda is critically needed to improve our understanding of the relationship between SIF and GPP across biomes, ecosystems, and even species. We advocate maintaining and upgrading current EC sites and adding ground-based SIF measurements to provide another scale of SIF observations. We also suggest