As the quality of machine translation improves, scholarly and pedagogical investigations on the roles and applications of machine translation in language classrooms continue to grow. The primary focus of such research is frequently concerned with the tool’s feasibility and usefulness in connection to language learning. However, in recent years, new proposals in the social sciences have emerged to redefine materials and objects beyond their long-held passive and static nature. Notable among them are a series of posthuman theories, including new materialism, which argue that the traditional understanding of matter and tools fails to adequately reflect their ontological position of acting and becoming by reducing them to objects for human use. This article proposes new onto-epistemologies of machine translation in language learning, based on recent theoretical discourse in new materialism research. To this goal, new materialist theories in regard to objects, machines, and devices are reviewed. Also, works of applied linguists, Alastair Pennycook, Kelleen Toohey, and Suresh Canagarajah, who have pioneered the new materialist turn, are introduced. In light of potentiality, performativity, and assemblage, this study presents different viewpoints on machine translation in language classrooms. Finally, this article calls for naturalistic and ethnographic research to advance a new materialist understanding of machine translation.