Accurate monitoring of Earth’s time-variable gravity field is essential for observing mass redistribution processes related to climate change, hydrology, cryosphere evolution, and solid Earth dynamics. Satellite gravimetry missions such as GRACE and GRACE Follow-On (GRACE-FO) have demonstrated the high potential of low–low satellite-to-satellite tracking (LL-SST) for recovering temporal gravity variations. However, the accuracy of gravity field recovery strongly depends on the quality of accelerometer (ACC) observations used for modeling non-gravitational forces. Conventional electrostatic accelerometers (EA) are affected by limitations in long-term stability and low-frequency noise, motivating the investigation of alternative sensor concepts and advanced data reconstruction approaches. This study investigates accelerometer data transplantation as a promising strategy for future GRACE-like satellite gravimetry missions. The proposed approach aims to reconstruct and transfer non-gravitational acceleration information between satellites, thereby reducing the dependence on high-performance accelerometers onboard every satellite of the constellation. Different mission scenarios are analyzed using closed-loop LL-SST simulations, including configurations equipped with classical electrostatic accelerometers, quantum-based Cold Atom Interferometry (CAI) accelerometers, and hybrid EA–CAI systems. Particular emphasis is placed on artificial intelligence (AI)-driven approaches, where machine learning models are used to learn cross-satellite relationships and reconstruct degraded or missing ACC observations. The developed models enable data-driven signal transplantation between satellites and support improved recovery of non-gravitational accelerations within the gravity field processing chain. The results demonstrate that AI-assisted ACC transplantation can significantly enhance accelerometer signal reconstruction and improve gravity field recovery performance. In particular, hybrid mission architectures combined with machine learning approaches provide a robust and cost-efficient alternative to fully instrumented dual-sensor configurations, while maintaining comparable recovery accuracy. Furthermore, the integration of quantum accelerometers shows strong potential for improving low-frequency stability, which is a critical requirement for next-generation satellite gravimetry missions. The presented work highlights the potential of combining AI-based signal reconstruction, accelerometer transplantation, and quantum sensing technologies to advance future satellite gravimetry missions. These developments support the design of more accurate, resilient, and cost-effective GRACE-like mission concepts for long-term monitoring of Earth’s dynamic gravity field.