Collaborative virtual reality (VR) is increasingly receiving attention, but the effects of context- specific variables and the interplay of telepresence, interactivity, and immersion as VR’s distinctive characteristics in such settings are little understood. Besides these three VR characteristics, we investigate in a quantitative study with 102 participants the influence of social presence, i.e. the sense of community; media naturalness, or the similarity of communication to face-to-face-interaction; and trust between users. Based on partial least squares structural equation modeling, we confirm the importance of interactivity and immersion, but not of telepresence. Moreover, we find that trust is essential for collaborative VR experiences, but social presence and media naturalness seem negligible. Finally, we show that immersion is a main driver of users’ intention to collaborate. Besides providing practitioners with insights for creating VR experiences, our study highlights that findings from research on individual VR use are not readily transferable to collaborative contexts.