Direct Air Capture (DAC) is a pivotal negative emission technology (NET), with adsorption-based DAC offering high energy efficiency and scalability. In this work, we leverage generative AI models trained on the CoreMOF database to design novel MOF structures with high CO₂ equilibrium uptake. These AI-generated candidates were evaluated via molecular-level simulations (GCMC/MD) to derive accurate isotherms. The obtained data bridges to process-level understanding through 1-D contactor simulations that capture dynamic adsorption behavior. Based on these dynamics, we calculate key techno-economic indicators such as levelized cost of CO₂ capture (LCOC), energy intensity (EI), and productivity. We further integrate these metrics into a climate-level optimization using two-stage stochastic programming to evaluate large-scale deployment impact. This study presents a complete multiscale framework, from material generation to climate systems modeling. The approach demonstrates how generative AI can accelerate DAC material discovery while providing predictive insights at process and policy levels. Overall, this work introduces a novel pathway for climate-material mapping via GenAI-driven DAC modeling.
