Under the dual background of global industrial chain intelligent restructuring and industrial-digital integration, artificial intelligence (AI) is reshaping the logic of corporate asset structure allocation and has become a core engine driving micro-level resource allocation. However, existing literature lacks systematic evidence on how AI affects corporate asset structure misallocation, especially the internal transmission mecha- nisms and heterogeneous conditions.

Using a sample of China’s A-share listed companies from 2009 to 2024, this paper empirically examines the impact of AI on corporate asset structure allocation. The core explanatory variable is the AI application level measured by text analysis of annual reports, and the dependent variable is asset structure misallocation measured by the marginal product of capital deviation method. The results show that AI significantly reduces corporate asset structure misallocation, with an economic magnitude of 6.38% of the sample mean, indicating that AI helps firms allocate assets more reasonably. Mechanism testing shows that AI alleviates asset structure misallocation through three pathways: suppressing strategic deviation, lowering agency costs, and enhancing digital technology application. Heterogeneity analysis reveals that the mitigating effect of AI on asset structure misallocation is more pronounced in firms located in regions with higher levels of industrial transformation and Fintech development, as well as in firms with higher digital infrastructure, more data assets, and higher analyst attention, and those characterized by “less talk but more action” in AI adoption.

The marginal contributions of this paper are as follows: First, it enriches the micro-level research on the economic consequences of AI by directly linking AI to corporate asset structure optimization. Second, it opens the black box of internal mechanisms from the perspectives of strategic deviation, agency costs, and digital technology application. Third, it offers nuanced heterogeneous evidence to guide targeted policies. Based on these findings, policy recommendations include that governments should build AI application promotion platforms and provide low-cost solutions for SMEs, that firms should proactively integrate AI into asset management and reduce strategic bias, and that policies should prioritize regions and firms with favorable conditions while helping laggards strengthen digital foundations.