Biogenic CO₂ emissions in cities are shaped by urban land cover change which can release carbon stocks, and, carbon sequestration by in situ vegetation. To date, these two processes have not been studied together and compared with transboundary fossil fuel-based CO₂ emissions of urban energy use. We leverage remote sensing and machine learning to quantify biogenic CO₂ emissions between 2006 and 2012, across 11 U.S cities, including central and suburban cities, in different climate zones. Results indicate that in situ carbon sequestration by greenery varied moderately across cities (−2.1 to −0.87 Mg CO₂ ha^–1 yr^–1), while emissions from the carbon stock change due to land conversion varied much more (−3.4 to 9.8 Mg CO₂ ha^–1 yr^–1), indicating that the latter dominates biogenic CO₂ emissions. Net biogenic CO₂ emissions were negative (carbon sink) in four cities, while large net positive emissions were present in rapidly growing suburbs. As a ratio of community-wide energy use for travel and buildings, biogenic CO₂ emissions were a small proportion in the core cities Denver (0.17%) and Minneapolis (0.33%) and as high as 38.2% in growing exurban communities. These results show that land cover change and greenery will be important policy levers in zero-carbon city planning.