Deploying seismic or infrasound arrays on the ground to probe a planet’s interior structure remains challenging in remote regions facing harsh surface conditions such as Venus with a surface temperature of 464°C. Fortunately, a fraction of the seismic energy transmits in the upper atmosphere as infrasound waves, i.e. low-frequency pressure perturbations (< 20Hz). On July 22, 2019, a heliotrope balloon, equipped with pressure sensors, was launched from the Johnson Valley, CA with the objective of capturing infrasound signals from the aftershock sequence of the 2019 Ridgecrest earthquake. At 16:27:36 UTC, the sound of a natural earthquake of Mw 4.2 was detected for the first time by a balloon platform. This observation offered the opportunity to attempt the first inversion of seismic velocities from the atmosphere. Shear velocities extracted by our analytical inversion method fell within a reasonable range from the values provided by regional tomographic models. While our analysis was limited by the observation’s low signal-to-noise ratio, future observations of seismic events from a network of balloons carrying multiple pressure sensors could provide excellent constraints on crustal properties. However, to build robust estimates of seismic properties, inversion procedures will have to account for uncertainties in terms of velocity models, source locations, and instrumental errors. In this contribution, we will discuss the current state of balloon-based observations, the sensitivity of the acoustic wavefield on subsurface properties, and perspectives on future inversions of seismically-induced acoustic data.