Climate change is raising the frequency of extreme heat events, creating growing challenges for agricultural production (OECD, 2023). While much research has studied the large-scale impacts of climate stressors on field crops and yields, similar evidence for livestock and dairy systems remains limited, especially regarding output quality. We link granular cow-level data with detailed weather conditions to examine the nonlinear effects of heat stress on milk yield, as well as fat and protein concentrations. Our dataset tracks more than 8 million dairy cows across the United States over a decade (2007-2016), providing multiple daily production snapshots throughout each lactation. Our measure of heat stress captures the amount of time spent at various intervals of the Temperature-Humidity Index (THI) during the seven days preceding each milk test day. We find that milk yield declines markedly when THI exceeds 70. In contrast, fat and protein content decline more gradually, beginning around a THI of 55, indicating that even moderate heat stress can harm dairy output quality. Our findings offer a more detailed understanding of how climate stressors affect dairy production and provide insights into strategies for adaptation to reduce future productivity losses.