Transparent electrodes are essential components in a huge variety of energy, lighting and heating devices with indium tin oxide being the most efficient and widely used so far. However, due the brittleness and scarcity of indium, an intensive research interest has emerged the last decade towards alternative transparent conductive materials (TCM). Metallic nanowire networks and especially silver nanowire (AgNW) networks appear to be one of the most promising among them, thanks to their excellent electrical and optical properties combined with their excellent mechanical performance and low cost fabrication. Despite the optimization of AgNW network fabrication methods and properties, there are still challenges to be tackled, in order to build a mature technology that can be successfully integrated into devices. The present PhD thesis focuses on the fundamental understanding of the physical phenomena that take place at both scales of nanowires and networks. Combining both experimental and modelling approaches, one of the main goals focuses on the origin of failure in AgNW networks during electrical stress. In situ measurements of the electrical resistance with a parallel recording of the spatial surface temperature by IR imaging are techniques that provide valuable information for the degradation mechanisms in a AgNW network. The simulation of the electrical distribution and power-induced heating offer a deeper understanding of the underlying physics and can be used to predict the networks electrical and heating performances. Moreover, an experimental study, conducted under simultaneous electrical and thermal stress, was useful for the successful integration of AgNW into devices. Based on these data, a physical model was proposed for the prediction of the time of failure, with its dependence with temperature, electrical current and AgNW network density (i.e. electrical resistance). Another crucial parameter that is investigated during this PhD thesis, is the presence of defects and the impact of networks non-homogeneity on the electrical distribution, the dynamics of failure and the surface temperature induced by Joule heating. Furthermore, the enhancement of the AgNW networks stability was successful with the network encapsulation by transparent, protective oxides developed in LMGP by Atmospheric Pressure Spatial Atomic Layer Deposition (AP-SALD). The AP-SALD is now an emerging open air, low-cost and scalable deposition method and the resulting AgNW-oxide composites retain an excellent flexibility. Finally, during the present PhD thesis, the integration of AgNW networks in devices was studied in the framework of several collaborations and projects with scientific teams of laboratories in Grenoble and elsewhere. Transparent heaters (TH) are in the core of our research and biomedical lab-on-a-chip devices were thoroughly studied. In addition, the increasing interest of the TCM community to the TH, lead our team to write and publish recently a review-article on the topic, including all different technologies and the physics related to Joule heating and the associated applications. Another example of studied integration is the use of the AgNW embedded in polymer substrates, as stretchable electrodes for electrostatic or piezoelectric generators in energy harvesting devices. In the framework of ANR projects with laboratory and industrial partners, we studied the use of AgNWs in cold field emission for miniaturized X-ray sources and as transparent, flexible electrodes in organic photovoltaic. To conclude, the AgNW networks optimized electrical, optical and mechanical properties and their combination with other transparent thin films or 2D materials are highly promising, as well as the growing industrial interest for their implementation. During this PhD Thesis, potential pathways for the stability enhancement and the improvement of the integration into devices have been traced, by means of both experimental and modelling approaches.