Atmospheric methane is the second most important greenhouse gas after carbon dioxide, and is responsible for about 20% of the global warming effect since pre-industrial times. Rice paddies are the largest anthropogenic methane source and produce 7-17% of atmospheric methane. Warm waterlogged soil and exuded nutrients from rice roots provide ideal conditions for methanogenesis in paddies with annual methane emissions of 25-100-million tonnes. This scenario will be exacerbated by an expansion in rice cultivation needed to meet the escalating demand for food in the coming decades. There is an urgent need to establish sustainable technologies for increasing rice production while reducing methane fluxes from rice paddies. However, ongoing efforts for methane mitigation in rice paddies are mainly based on farming practices and measures that are difficult to implement. Despite proposed strategies to increase rice productivity and reduce methane emissions, no high-starch low-methane-emission rice has been developed. Here we show that the addition of a single transcription factor gene, barley SUSIBA2 (refs 7, 8), conferred a shift of carbon flux to SUSIBA2 rice, favouring the allocation of photosynthates to aboveground biomass over allocation to roots. The altered allocation resulted in an increased biomass and starch content in the seeds and stems, and suppressed methanogenesis, possibly through a reduction in root exudates. Three-year field trials in China demonstrated that the cultivation of SUSIBA2 rice was associated with a significant reduction in methane emissions and a decrease in rhizospheric methanogen levels. SUSIBA2 rice offers a sustainable means of providing increased starch content for food production while reducing greenhouse gas emissions from rice cultivation. Approaches to increase rice productivity and reduce methane emissions as seen in SUSIBA2 rice may be particularly beneficial in a future climate with rising temperatures resulting in increased methane emissions from paddies.