Telegram has become a preferred platform for far-right activism, conspiracy theories, political propaganda, and misinformation, each with its own target audience. This study investigates the application of multilanguage machine learning techniques for extracting topics from political content on Telegram channels. This research improves the understanding of political information and narratives in different languages on Telegram over time, contributing to the study of digital communication and information warfare within the content of the Telegram channels sphere. Through the exploitation of an extensive dataset on the subject, this work contributes to the state-of-the-art of topic analysis by providing information on prevailing discourses, including the escalating of armed conflicts between Russian and Ukraine, and political tensions in Europe and the USA.