This study conducts a comparative analysis on the use of multimodal interactional metadiscourse by U.S. and Chinese companies on Twitter. Six hundred tweets generated by twelve U.S. and Chinese online retailing, technology, energy, and automobile companies were analyzed. The results indicate that corporate tweets are highly interpersonal in nature, rich in interactional metadiscourse. Attitude markers, self-mentions, and a series of engagement markers are particularly prominent. Various non-textual devices, such as pictures, videos, @-mentions, hashtags (#), and hyperlinks, are co-deployed with textual elements to construct interpersonal meanings. Overall U.S. companies used more engagement markers, while Chinese companies employed more stance markers, with other salient differences observed in different subcategories. These variances can be attributed to various sociocultural and contextual factors. The study contributes to the still scarce research on multimodal metadiscourse in online business discourse and uncovers the influence of sociocultural and contextual factors on the use of interpersonal resources in digital business genres.