Abstract: The high-investment nature of the artificial intelligence industry has prompted countries to intensively introduce government subsidy policies, yet the lag of existing international rules has triggered numerous legal disputes, making it imperative to strengthen research on the regulation under international economic law. Taking government subsidies for AI enterprises as the research object, this paper systematically analyzes relevant international economic law issues. In terms of the rule framework, the Agreement on Subsidies and Countervailing Measures of the World Trade Organization contains ambiguous provisions on the characterization of subsidies for AI basic research and funding for computing infrastructure, and there are conflicts in the cross-application with digital trade rules such as the General Agreement on Trade in Services. In terms of practical disputes, the discriminatory subsidy clauses implemented by Europe and the United States through initiatives like the Digital Europe Programme and the CHIPS and Science Act have formed a game with the inclusive support provided by China’s AI funds, highlighting the risk of global industrial chain fragmentation. Drawing from typical cases such as the EU’s review of R&D funding for ASML, and the disputes surrounding China’s “Eastern Data and Western Computing” project, this paper proposes the development trends of government subsidies for AI enterprises and their regulatory optimization paths. Firstly, global government subsidies for AI enterprises need to strike a balance between innovation incentives and trade fairness. The focus of future rule games will be on the definition of infrastructure subsidies and the boundary of application of the “national security exception”, so as to promote the modernization of World Trade Organization rules and the coordination of regional agreements. Secondly, the Chinese government should optimize the structure of government subsidies, strengthen multilateral coordination, and achieve breakthroughs in technical standards; AI enterprises need to establish diversified supply chain systems and compliance review mechanisms.