Here, we report the morphological and electrical properties of self-assembled silicon nanowires networks, also called Si nanonets. At the macroscopic scale, the nanonets involve several millions of nanowires. So, the observed properties should result from large scale statistical averaging, minimizing thus the discrepancies that occur from one nanowire to another. Using a standard filtration procedure, the so-obtained Si nanonets are highly reproducible in terms of their morphology, with a Si nanowire density precisely controlled during the nanonet elaboration. In contrast to individual Si nanowires, the electrical properties of Si nanonets are highly consistent, as demonstrated here by the similar electrical properties obtained in hundreds of Si nanonet-based devices. The evolution of the Si nanonet conductance with Si nanowire density demonstrates that Si nanonets behave like standard percolating media despite the presence of numerous nanowire-nanowire intersecting junctions into the nanonets and the native oxide shell surrounding the Si nanowires. Moreover, when silicon oxidation is prevented or controlled, the electrical properties of Si nanonets are stable over many months. As a consequence, Si nanowire-based nanonets constitute a promising flexible material with stable and reproducible electrical properties at the macroscopic scale while being composed of nanoscale components, which confirms the Si nanonet potential for a wide range of applications including flexible electronic, sensing and photovoltaic applications.