AbstractMany of the irrigated spring wheat regions in the world are also regions with high poverty. The impacts of temperature increase on wheat yield in regions of high poverty are uncertain. A grain yield–temperature response function combined with a quantification of model uncertainty was constructed using a multimodel ensemble from two key irrigated spring wheat areas (India and Sudan) and applied to all irrigated spring wheat regions in the world. Southern Indian and southern Pakistani wheat‐growing regions with large yield reductions from increasing temperatures coincided with high poverty headcounts, indicating these areas as future food security ‘hot spots’. The multimodel simulations produced a linear absolute decline of yields with increasing temperature, with uncertainty varying with reference temperature at a location. As a consequence of the linear absolute yield decline, the relative yield reductions are larger in low‐yielding environments (e.g., high reference temperature areas in southern India, southern Pakistan and all Sudan wheat‐growing regions) and farmers in these regions will be hit hardest by increasing temperatures. However, as absolute yield declines are about the same in low‐ and high‐yielding regions, the contributed deficit to national production caused by increasing temperatures is higher in high‐yielding environments (e.g., northern India) because these environments contribute more to national wheat production. Although Sudan could potentially grow more wheat if irrigation is available, grain yields would be low due to high reference temperatures, with future increases in temperature further limiting production.