The TED talk (www.ted.com) is a hybrid popularizing genre empowered by contemporary digital technologies in which different semiotic modes feature prominently and which is being extensively used in educational settings. This study is based on and further develops research on co-speech gestures in a selection of such talks from various knowledge domains, so as to shed light and raise awareness on the orchestration of different modal resources therein and as a way of contributing to the development of multimodal literacy in an ever changing educational landscape. Data description is based on multimodal transcription through an integrated method, which makes it possible to advance hypotheses about the interpretation of gestures in different contexts. The qualitative analysis will show various ways in which speech-synchronized gestures in the talks can contribute different (also simultaneous) ideational, interpersonal and textual meanings or metafunctions, especially when considered from a more global analytical perspective, viz. as repeated similar patterns over discourse chunks. Reference will also be made to Multimodal Critical Discourse Analysis as an overarching framework, as some of the patterns appear to have the potential not only to enhance cohesion but also to subtly emphasize emotional and value-laden meanings, thus pushing the highly persuasive discourse of this genre of talks forward.