Artificial water channels (AWCs) and their natural aquaporin counterparts selectively transport water. They represent a tremendous source of inspiration to devise biomimetic membranes for several applications, including desalination. They contain variable water-channel constructs with adaptative architectures and morphologies. Herein, we critically discuss the structural details that can impact the performances of biomimetic I quartets, obtained via adaptive self-assembly of alkylureido-ethylimidazoles HC4-HC18 in bilayer or polyamide (PA) membranes. We first explore the performances in bilayer membranes, identifying that hydrophobicity is an essential key parameter to increase water permeability. We compare various I quartets with different hydrophobic tails (from HC4 to HC18), and we reveal that a huge increase in single-channel water permeability, from 104 to 107 water molecules/s/channel, is obtained by increasing the size of the alkyl tail. Quantitative assessment of AWC-PA membranes shows that water permeability increases roughly from 2.09 to 3.85 L m-2 h-1 bar-1, for HC4 and HC6 reverse osmosis membranes, respectively, while maintaining excellent NaCl rejection (99.25-99.51%). Meanwhile, comparable HC8 loading induces a drop of performance reminiscent of a defective membrane formation. We show that the production of nanoscale sponge-like water channels can be obtained with insoluble, low soluble, and low dispersed AWCs, explaining the observed subpar performance. We conclude that optimal solubility enabling breakthrough performance must be considered to not only maximize the inclusion and the stability in the bilayer membranes but also achieve an effective homogeneous distribution of percolated particles that minimizes the defects in hybrid polyamide membranes.