The increasingly interconnected global food system is becoming more vulnerable to production shocks owing to increasing global mean temperatures and more frequent climate extremes. Little is known, however, about the actual risks of multiple breadbasket failure due to extreme weather events. Motivated by the Paris Climate Agreement, this paper quantifies spatial risks to global agriculture in 1.5 and 2°C warmer worlds. This paper focuses on climate risks posed to three major crops-wheat, soybean and maize-in five major global food producing areas. Climate data from the atmosphere-only HadAM3P model as part of the "Half a degree Additional warming, Prognosis and Projected Impacts" (HAPPI) experiment are used to analyse the risks of climatic extreme events. Using the copula methodology, the risks of simultaneous crop failure in multiple breadbaskets are investigated. Projected losses do not scale linearly with global warming increases between 1.5 and 2°C Global Mean Temperature (GMT). In general, whilst the differences in yield at 1.5 versus 2°C are significant they are not as large as the difference between 1.5°C and the historical baseline which corresponds to 0.85°C above pre-industrial GMT. Risks of simultaneous crop failure, however, do increase disproportionately between 1.5 and 2°C, so surpassing the 1.5°C threshold will represent a threat to global food security. For maize, risks of multiple breadbasket failures increase the most, from 6% to 40% at 1.5 to 54% at 2°C warming. In relative terms, the highest simultaneous climate risk increase between the two warming scenarios was found for wheat (40%), followed by maize (35%) and soybean (23%). Looking at the impacts on agricultural production, we show that limiting global warming to 1.5°C would avoid production losses of up to 2 753 million (161 000, 265 000) tonnes maize (wheat, soybean) in the global breadbaskets and would reduce the risk of simultaneous crop failure by 26%, 28% and 19% respectively.