Could a widespread proliferation of ridesharing services mitigate or exacerbate the carbon footprint of urban passenger transport? Despite having profound policy implications, this question has not yet been answered in the literature. This paper examines that impact ex-ante, by simulating the aggregate travel demand, the choice of transport mode and the resulting CO2 emissions in 247 cities between 2015 and 2050. We find that if ridesharing services receive substantial policy support, CO2 emissions from passenger transport in 2050 will be on average 6.3% lower than their reference level. However, we show that this finding differs widely across cities. The paper identifies the reasons for this variation and the policies that are socially desirable in a given city, conditional on its characteristics.