While there has been a scholarly call for the incorporation of digital multimodal composing (DMC) in L2 pedagogy over the past decade, the ideologies of teachers and students about using DMC and the pertinent impact on students’ investment in language learning remain underexplored. This paper is concerned with the experiences of a group of English teachers and students when they participated in a DMC program that engaged students with video production in a university-based English course in China. Relying on multiple sources of data, the study revealed that teachers and students had different, and sometimes even contrasting, ideologies over the nature of language, the role of teachers, and the legitimate evidence of learning during DMC. A framework is proposed to illustrate how the different and contrasting ideologies created unintended barriers to students’ investment in English learning at both micro and macro levels.