A blessing or a curse? The impact of artificial intelligence innovation on corporate energy efficiency
This study examines how artificial intelligence (AI) innovation influences corporate energy efficiency and the underlying mechanisms. We develop a general equilibrium model incorporating energy input and endogenous AI technology to explain how AI innovation improves energy efficiency through production process optimization. Using a deep-learning-based text identification, we develop a firm-level measure of AI innovation for Chinese listed companies over the period 2008–2022. Our empirical results show that AI in…

In brief
Researchers built a general equilibrium model with energy input and endogenous AI technology, then measured firm-level AI innovation for Chinese listed companies with deep-learning-based text identification. Their data cover 2008 to 2022, and they found that firms with more AI innovation used energy more efficiently.
The finding matters because it ties AI to corporate energy use through a measurable operational channel rather than through general claims about technology. The abstract reports an association strengthened by certain conditions, and it does not show whether the result holds beyond the sampled Chinese listed firms or reflects total energy demand.
Main points
- The study uses a firm-level AI innovation measure built through deep-learning-based text identification.
- The sample is Chinese listed companies over 2008 to 2022.
- Mechanism analysis is consistent with production process optimization as a transmission channel.
- The effect is stronger in key technological sectors and capital-intensive industries.
- Disruptive AI innovation has a larger effect than general AI innovation.
The gain
Across Chinese listed companies from 2008 to 2022, greater AI innovation was associated with significantly better corporate energy efficiency, with production process optimization as a likely channel.
The rundown
The authors developed a theoretical model before testing it empirically. It includes energy as a production input and treats AI technology as endogenous, which they use to explain how AI innovation can raise energy efficiency by optimizing production processes.
The empirical analysis found heterogeneous effects. Gains were stronger for firms in key technological sectors and capital-intensive industries, and disruptive AI innovation produced a larger effect than general AI innovation.
Context also mattered. Local governments' attention to AI development and intra-firm data sharing strengthened the efficiency gains, and the authors frame the result as AI-driven innovation working through organizational and data-governance mechanisms.
Sources
- Peer-reviewedJournal of Environmental Management2026-10-19
ace


The debate