Reconciling various geophysical models generated from multiphysics data is critical to fully understanding geothermal systems. In particular, inverted density models from airborne gravity gradiometry (AGG) survey data yield information about faulting and fracturing, whilst inverted resistivity models produced by the magnetotelluric (MT) method image the fluids and geothermal source rock of the system. Joint Gramian inversion of multiphysics data can deliver models with shared structure and correlated physical properties or their transforms, facilitating a better and more consistent geological picture of the system. To meet this challenge, we applied 3D joint Gramian inversion of AGG and MT data to study Japan's geothermal resources. Japan Organization for Metals and Energy Security (former Japan Oil, Gas and Metals National Corporation, JOGMEC), Idemitsu Kosan Co., Ltd. (Idemitsu), and TechnoImaging conducted extensive geophysical studies of a geothermal field in Japan, including AGG and MT surveys. The project aimed to study the location and structure of geothermal energy sources in the survey area. The observed AGG and MT data were analyzed separately and jointly using 3D inversion methods. For joint inversion, we used the approach based on Gramian constraints, which enforce the correlation between the different physical parameters of the inverse models, or their transforms, including gradients, thus ensuring that the inversion produces a consistent image of the subsurface geological formation. The comparison of results of the standalone and joint inversions demonstrates the effectiveness of the developed method for geothermal resource exploration.