To have a reasonable assessment of the impact of RE beams in ITER, it is necessary to understand how the beam interacts with the background MHD and the pathways to termination and possible reformation. In this regard, we investigate the termination of an ITER RE beam via 3D non-linear MHD simulations using the JOREK code [1, 2]. The RE fluid model [3] has been used for this purpose which includes the effects of partially-ionized impurities [4]. The specific scenario considered here corresponds to a plateau-phase RE beam in an elongated cold plasma undergoing a vertical motion towards the wall. The current profile is mildly hollow with an edge safety factor close to 2. Both the factors (hollow current profile and q95 ~ 2) are observed to contribute to a strong destabilization, leading to a stochastic magnetic field and subsequent near-complete loss of RE beam current. The observation is relatively robust over a range of parameters explored in the study. At high electrical resistivity, MHD dynamics are markedly different, leading to stronger stochastization and lower remnant REs. This suggests that large resistivity of low temperature plasmas (that is typical of plasmas with RE beams), is likely an important factor in the robust observation of benign termination observed is several experiments [5, 6]. RE load distributions on the first wall and the propensity for remnant REs to re-avalanche will also be discussed. References [1] G.T.A. Huijsmans et al., Nucl. Fusion 47.7, 659 (2007). [2] M. Hoelzl et al., Nucl. Fusion 61, 065001 (2021). [3] V. Bandaru et al., Phys. Rev. E 99, 063317 (2019). [4] L. Hesslow et al., Nucl. Fusion 59, 084004 (2019). [5] C. Reux et al., Phys. Rev. Lett. 126, 175001 (2021). [6] Paz-Soldan et al., Nucl. Fusion 61, 116058 (2021).