This chapter probes into China's agile approach to artificial intelligence (AI) governance that aims to balance AI risk mitigation and fostering innovation. By adopting a point–line–plane analytical strategy, this study first maps the stakeholders in China's AI policymaking (the points), then analyses interactions between stakeholders (the lines), and finally depicts the larger picture of the governance system as a whole (the plane). The findings suggest that China's AI governance involves predominantly six different groups of stakeholders, whose interests converge, compete, and collide respectively in the situations of boosting China's AI capabilities, creating a Responsible AI agenda, and governing generative AI. This bespeaks a governance paradigm shift towards multistakeholderism in which trust between stakeholders is essential for more efficient cooperation and determines if this agile approach remains immature or a sustainable model worthy of following.