We pair a firm-level panel of annual industrial output with a fine-scale daily weather data set, to estimate the responses of industrial output to temperature changes in China. We have four primary findings. First, industrial output is nonlinear in temperature changes. With seasonal average temperatures as temperature variables, output responds positively to higher spring temperatures and negatively to elevated summer temperatures. With temperature bins as temperature variables, output increases linearly with temperature up to 21–24 °C, and then declines sharply at higher temperatures.