In　recent　years,　data　has　increasingly　become　the　“new　oil”　for　21st-century　economic　development.　However,　there　is　still　a　gap　in　how　the　development　of　big　data　promotes　the　improvement　of　urban　carbon　unlocking　efficiency　(UCUE).　Utilizing　advanced　double　machine　learning　(DML)　methods,　and　treating　the　big　data　comprehensive　pilot　zone　(BDCPZ)　as　a　quasi-natural　experiment,　we　employ　panel　data　from　282　Chinese　cities　spanning　2011　to　2022　to　study　the　impact　of　big　data　policies　on　UCUE　and　its　mechanisms.　The　study　finds　that:　(1)　Big　data　policies　significantly　enhance　carbon　unlocking　efficiency,　and　their　importance　in　carbon　unlocking　is　confirmed　even　when　alternative　machine　learning　models　are　used.(2)　Regarding　the　mechanisms,　big　data　policies　improve　carbon　unlocking　efficiency　through　three　pathways:　government　modernization,　enterprise　intelligent　development,　and　economic　transformation.(3)　Heterogeneity　analysis　reveals　that　the　carbon　unlocking　benefits　of　big　data　policies　are　more　pronounced　in　large　cities,　old　industrial　base　cities,　digital　economy　dividend　cities　and　key　environmental　protection　cities.　We　also　provide　insights　for　strengthening　the　construction　of　big　data,　alleviating　carbon　emission　pressures,　and　achieving　the　goals　of　“dual　carbon”.　©　2024　Elsevier　Ltd