Search for a command to run...
Digital technologies are often presented as great sources of efficiency. They however also cause significant environmental damages. Rebound effect is defined as efficiency gains leading to increased resource use. Its role in the environmental impact of digital technologies remains unclear because it is unknown whether the environmental impacts raise because of or in spite of the efficiency gains. This dynamics is often explored by relating efficiency gains to changes in demand following an economic perspective.Stakeholders leverage energy efficiency along other technical changes as parts of their development strategies. Accounting for the sociotechnical contexts of technological developments is therefore instrumental to understand environmentalimpacts dynamics. In this paper, we analyze the sociotechnical context of two case studies: the roll-out of 5G in France, and, deep learning model training. We identify technical characteristics and actors intentions to clarify the role energy efficiency plays in these technological developments and their (environmental) consequences.In both cases, major actors openly show intent towards activity growth. Energy efficiency serves as one of the levers to achieve their goal. The concept of rebound effect thus seems inadequate to capture the dynamics at play in our case studies. Indeed, energy efficiency is not used (principally) to reduce impact, but rather to allow for the creation of new uses.We believe that similar conclusions could be obtained on numerous other aspects of digitalization by using our methodological framework. Our results suggest rethinking the framing of energy efficiency to better approach the environmental consequences of digitalization.