AI-driven echelon utilization of retired electric vehicle batteries and their life cycle carbon mitigation potential

Abstract The rapid retirement of electric vehicle (EV) batteries presents both a challenge of resource management and an opportunity for carbon mitigation. An Artificial Intelligence (AI)-driven framework is introduced that couples high-accuracy capacity forecasting with a dynamic life cycle carbon model (LCCO 2 ), enabling intelligent end-of-life decision-making and large-scale repurposing. Analysis of real-world field data from commercial EVs shows that data-driven retirement strategies extend battery service life beyond conventional fixed thresholds, increase usable energy output, and improve classification accuracy of safety margins. Integration into the LCCO 2 model indicates reductions of 14.3% in average life cycle carbon intensity, alongside substantial per-battery CO 2 savings, without compromising operational safety. At scale, projections suggest that second-life deployment of retired EV batteries could deliver tens of tens of millions of metric tons of CO 2 mitigated annually in China and the European Union by mid-century, thereby contributing directly to both regions’ carbon neutrality pathways. These findings highlight a scalable strategy that unites predictive modelling with carbon accounting, advancing the circular economy of EV batteries. Graphical Abstract