Existing empirical studies suggest that statistical laws of information avalanches in social media lack robustness across systems. This paper analyzes nearly one billion time-stamped events from various online platforms (Telegram, Twitter, Weibo, etc.) over extended periods (more than ten years). It demonstrates that information propagation in social media is a universal and critical process, exhibiting identical macroscopic patterns across platforms regardless of system specifics. Power-law distributions and hyperscaling relations characterizing avalanche size and duration support the critical behavior. Statistical testing reveals a mixture of simple and complex contagion, with the complexity correlated to the information's semantic content.
