This study explores whether teacher performance trajectory over time differs by school-poverty settings. Focusing on elementary school mathematics teachers in North Carolina and Florida, we find no systematic relationship between school student poverty rates and teacher performance trajectories. In both high- (=60% free/reduced-price lunch [FRPL]) and lower-poverty (<60% FRPL) schools, teacher performance improves the fastest in the first 5 years and then flattens out in years 5 to 10. Teacher performance growth resumes between year 10 and 15 in North Carolina but remains flat in Florida. In both school-poverty settings, there is a significant variation in teacher performance trajectories. Among novice and early-career teachers, the fastest-growing teachers (75th percentile) improve by 0.04 standard deviations more in student gain scores annually than slower teachers (25th percentile). In both school settings, novice teachers who started with low effectiveness also grew at a slower rate in the next 5 years than novice teachers with higher initial effectiveness. Our findings suggest that the lack of productivity "return" to experience in high-poverty schools reported in the literature is unlikely to be the result of differential teacher learning in high- and lower-poverty schools.