The targeted fiscal expenditures on artificial intelligence (AI) represent a key governmental strategy to support the widespread adoption of AI technology in manufacturing, which is expected to drive AI development and facilitate the transformation and upgrading of the industry. However, despite the rapid advancement of intelligent manufacturing and the increasing application of industrial robots, China’s share of global manufacturing value-added has not shown a corresponding increase. Moreover, the effectiveness of fiscal expenditures in this context remains uncertain. These issues attract significant academic attention.

Using the panel data from Chinese cities between 2008 and 2019 and integrating both theoretical and empirical perspectives, this paper examines the impact of AI on manufacturing upgrading. The findings reveal that AI significantly enhances manufacturing upgrading. Specifically, a 1% increase in AI adoption corresponds to a 9.556% increase in the manufacturing upgrading rate. This effect is particularly pronounced in cities with a higher level of digital economy development, peripheral cities, and during periods of manufacturing growth. Mechanism testing indicates that AI drives manufacturing upgrading by enhancing labor force aggregation, boosting R&D innovation capabilities, and optimizing the factor allocation structure. Further research shows that fiscal expenditures positively moderate the impact of AI on manufacturing upgrading, especially in areas related to education and technology spending. However, fiscal expenditures on enterprise transformation and basic construction do not have a significant impact. This suggests that a lack of policy resultant force in fiscal expenditures aimed at promoting AI development may contribute to the asynchronous growth between AI applications and the manufacturing industry.

The marginal contributions of this paper are as follows: First, from the perspective of advanced technology and computing power enabling manufacturing upgrading, it constructs a measurement index for manufacturing upgrading to assess the impact of AI, aligning more closely with the evolving dynamics of modern manufacturing development. Second, after establishing that AI can promote manufacturing upgrading, it examines the heterogeneity of AI’s impact across three dimensions: digital economy development level, city type, and industry cycle, thereby broadening the scope of existing research. Third, it explores the mechanisms through which AI affects manufacturing upgrading, offering a novel theoretical framework for understanding the interaction between AI and manufacturing upgrading. Fourth, it identifies that current fiscal expenditure policies in China have not yet formed a resultant force to enhance AI’s role in promoting manufacturing upgrading, providing valuable insights for optimizing fiscal expenditure policies to support AI technology development.