Divergent paths: A machine learning analysis of post-pandemic decarbonization trajectories and the global climate policy imperative

This study analyzes global decarbonization pathways using trend and machine learning–based clustering of CO2 emissions from 20 major economies responsible for over 80 % of global emissions. The model used emissions intensity (CO2/TJ) and absolute levels in million tons (Mt) per annum. The analysis identified four archetypes: Sustainable Performers, Scale Challenges, Efficiency Laggards, and Critical Emitters. The pandemic was a reflexive experiment that exposed the fragility of decarbonization progress, signaled by differing recovery strategies. Gains from green stimulus investments in countries like Germany, the U.K., and the U.S. were counterbalanced by much sharper post-pandemic rebounds in China (+407 Mt/year), India, and Nigeria (+7.1 CO2/TJ/year) under fossil-intensive recoveries driven by energy security concerns. Scenario modeling shows that only the Net Zero pathway with an 83 % emissions reduction by 2050 adheres to the 1.5 °C climate target, while an imminently projected business-as-usual growth would overshoot global emissions of 53–55 Gt by 2030 and breach the 2 °C limit. There are still many relevant structural barriers left unattended: policy inertia, fragmented recovery policies, and entrenched dependence on fossil fuels. A coalition-based global approach is the only way forward and should include binding emissions targets, scaled-up clean energy finance, and promoting equity in low-carbon infrastructure access. Projections based on these clusters indicate a most-likely trajectory of ∼40 Gt CO2/year by 2050, a projection insufficient for either 1.5 °C or 2.0 °C pathways. If left unaddressed, global decarbonization will fracture along clusters, structural and geopolitical lines, derailing climate goals while deepening global inequality. This analysis offers a reproducible framework to classify national transitions and guide equitable, cluster-based climate action.