Tritium permeation through the intermediate heat exchanger of the HYLIFE-II inertial fusion design concept is evaluated for routine operating conditions. The permeation process is modelled using the Lewis analogy combined with surface recombination. It is demonstrated that at very low driving potentials, permeation becomes proportional to the first power of the driving potential. The model predicts that under anticipated conditions the primary cooling loop will pass about 6% of the tritium entering it to the intermediate coolant. Possible approaches to reducing tritium permeation are explored. Permeation is limited by turbulent diffusion transport through the molten salt. Hence, surface barriers with impedance factors typical of present technology can do very little to reduce permeation. Low Flibe viscosity is desirable. An efficient tritium removal system operating on the Flibe before it gets to the intermediate heat exchanger is required. Needs for further research are highlighted.