These kinds of popsci graphs are quite misleading. Unlike with the “Spanish”* bubble, the “Chinese” and “Arabic” bubbles contain many mutually unintelligible languages. Though to be fair, many (most?) of the people in the “Arabic” bubble can speak/read Modern Standard Arabic as a lingua franca, and virtually everyone in China can read Standard Chinese.
*Also, Dude, Spanish is not the preferred nomenclature. Castilian, please.