The transition regime between giant planets (GPs) and brown dwarfs (BDs) is still an open subject of study in exoplanetary science. A complete understanding of the population of long-period GPs and BDs would be pivotal in understanding this topic, but the number of such objects with precisely measured orbital and physical parameters is still small. Moreover, their dynamical influence on smaller companions in inner orbits is still unclear. Within the GAPS programme, we aim to confirm and characterize sub-stellar companion candidates from Gaia DR3, and to study the potential presence of additional lower-mass planets in their systems. We present the results of an intensive high-precision radial velocity (RV) monitoring of HD128717, which hosts the astrometric candidate Gaia-ASOI-009. We used the HARPS-N spectrograph at TNG to collect a high-cadence RV time series of the target. We used Markov chain Monte Carlo (MCMC) analyses to refine the Gaia DR3 orbital solution of the companion and, finally, performed a combined model of RV and proper motion anomaly (PMa) to derive the complete 3-D orbit of the companion. We also ran a suite of numerical simulations to confirm our results. We confirm the sub-stellar nature of Gaia-ASOI-009, i.e. Gaia-6B: from the combined RV+PMa fit, we confirm that it is a high-eccentricity low-mass brown dwarf with P=9.37^+0.06^_-0.05_yr, M=19.8+/-0.5M_J_, e=0.85, i=130deg. The derived orbital solution differs significantly from the one published in Gaia DR3. Through a series of dedicated simulations, we demonstrate that this discrepancy arises from a degeneracy in the Gaia DR3 astrometric solution. Specifically, the combination of Gaia-6B long orbital period and high eccentricity, both poorly constrained by the limited timespan of DR3, led to an incorrect solution characterized by a shorter period and lower eccentricity. Finally, we found no evidence of other companions in the system, neither inner nor outer, and thus the origin of the high-eccentricity of Gaia-6B remains unclear.