Spectrum coexistence between terrestrial Next-G cellular networks and space-borne remote sensing (RS) is now gaining attention. One major question is how this would impact RS equipment. In this study, we develop a framework based on stochastic geometry to evaluate the statistical characteristics of radio frequency interference (RFI) originating from a large-scale terrestrial Next-G network operating in the same frequency band as an RS satellite. For illustration, we consider a network operating in the restricted L-band (1400-1427 MHz) with NASA's Soil Moisture Active Passive (SMAP) satellite, which is one of the latest RS satellites active in this band. We use the Thomas Cluster Process (TCP) to model RFI from clusters of cellular base stations on SMAP's antenna's main- and side-lobes. We show that a large number of active clusters can operate in the restricted L-band without compromising SMAP's mission if they avoid interfering with the main-lobe of its antenna. This is possible thanks to SMAP's extremely low side-lobe antenna gains.