The paper tries to test the thesis of increased "de-middledization" using international comparative data provided by Eurostat for 28 European Union countries and 3 of their neighbours (i.e., Switzerland, Norway, and Iceland) for the period from 2003 to 2014. Firstly, we examine the size of the middle class and changes within the time period based on the European Union Statistics on Income and Living Conditions (EU-SILC) dataset. The topic is closely connected to discussion in evolutionary economics as provided by Schumpeter’s (2000) idea of “creative destruction” which is led by the assumption of a fragile balance between births and deaths of firms, jobs, occupations, and economic sectors in the course of economic development. Of course, there is already some lively discussion on automation and digitalization as such (Acemoglu and Restrepo 2018; Autor 2015), but less research focusses explicitly at the middle classes. We find evidence of a negative link between digitalization and the size of the middle class, but the result is highly sensitive to the definition of middle class as the dependent variable. While there appears to be only a limited effect of the information and communication technology (ICT) contribution on gross domestic product (GDP) and total employment on the size of the middle class, the evidence is highly significant with respect to the effect of innovation in the digital sector as measured by ICT-related patents on the size of the middle class. One explanation for this might be ICT-related innovation benefited not only by those in ICT sectors, but also a wider array of occupations. Therefore, operationalizing the digitalization indicator by examining the contribution of the ICT sector to GDP and total employment could be misleading. With this paper, we have provided only the first test of the relationship between digitalization and the size of the middle class.