Digital Transformation, Carbon Disclosure, and Green Innovation in Energy Transition: Firm-Level Evidence from China’s Energy Sector

China’s dual-carbon goals have made understanding how digital transformation shapes corporate carbon performance a policy priority, and the rapid expansion of firm-reported emissions disclosure since 2020 now makes such inquiry feasible at scale. Drawing on CSMAR financial data and CNRDS digital-transformation keyword frequencies for 914 A-share energy firms over 2020–2025, we combine panel fixed-effects models, a placebo-validated triple-difference design, disclosure-choice models with a Heckman selection correction, and event studies of two policy milestones. We find that digital transformation robustly stimulates green-patent applications and significantly raises the likelihood of carbon disclosure in the cross section, identifying green-patenting and disclosure as the two margins on which digitalization is already measurable. The direct association with disclosed carbon intensity remains statistically insignificant in the selection-corrected model and on the directly disclosed subsample, indicating that emission reductions are not yet measurable at the firm level. A sharp but short-lived emission-reduction pulse among digitalized high-carbon firms in 2021 is consistent with but does not confirm a Porter-type mechanism. Investors nonetheless priced transition-regulatory exposure: ETS-covered firms earned significantly negative abnormal returns around both policy milestones. Thus, digitalization in China’s energy sector is already visible in what firms invent and disclose, with measurable emission and valuation effects likely to follow as disclosure mandates and carbon pricing mature.