Sustainable digital infrastructure and firm-level energy consumption: Evidence from China’s manufacturing sector

With the rapid advancement of artificial intelligence (AI) and the accelerating low-carbon transition, sustainable digital infrastructure has emerged as a critical physical foundation for improving energy management efficiency. Using panel data from manufacturing firms listed on the Shanghai and Shenzhen A-share markets in China over the period 2012-2024, this study employs a generalized continuous difference-in-differences (DID) model to systematically examine the impact of sustainable digital infrastructure on corporate energy consumption. The results show that sustainable digital infrastructure significantly reduces firms’ energy consumption. Mechanism analysis further reveals that this effect operates through two distinct channels: the promotion of disruptive green innovation and the enhancement of AI innovation. Heterogeneity analysis indicates that the energy-saving effect is highly context-dependent. It is more pronounced among firms located in regions benefiting from cross-regional computing power reallocation (eastern regions), and among firms characterized by lower climate risk exposure, lower levels of digital washing, and stronger absorptive capacity. In contrast, the effect is relatively weaker under the opposite conditions. Overall, this study extends the literature on digital infrastructure and corporate energy management by providing micro-level evidence on the energy-saving effects of sustainable digital infrastructure. It also offers policy-relevant insights for emerging economies seeking to advance low-carbon transitions through the development of green digital infrastructure.