With the increasing concerns of global climate change, clean energy development is regarded as one of the most important measures to mitigate CO2 emissions. However, existing studies pay little attention to the spatial spillover effect of clean energy development on CO2 emissions. Using a provincial-level panel data set during 1997-2017 and a spatial Durbin model, this is the first study to investigate the effect of clean energy develop-ment measured by the share of clean energy generation in total electricity generation on CO2 emissions in China. The results show that clean energy development causes less CO2 emissions in the local region but more CO2 emissions in spatially related regions. This finding is reinforced through a series of robustness checks. The heterogeneity analysis indicates that the local CO2 emission reduction effect of clean energy development is greater in the period of 2013-2017, electricity-poor regions, and low-carbon pilot regions; meanwhile, electricity-poor regions and low-carbon pilot regions suffer from a more adverse impact of clean energy devel-opment in spatially related regions on local CO2 emission reduction in the long run. Furthermore, the results of mechanism analysis suggest that the fossil energy saved by the usage of clean energy from the power generation sector in the local region flows into spatially related regions to crowd out clean energy consumption in these regions. Therefore, the overall effectiveness of CO2 emission reduction effort in clean energy development is partly undermined by the CO2 transfer effect. This paper provides a novel perspective to understand the effect of clean energy development on CO2 emission reduction and helps to promote inter-regional cooperative CO2 emission reduction within a country.