In human males, testosterone (T) decreases in the period following the birth of offspring. This decline has been widely interpreted as a facultative neuroendocrine response that facilitates parenting effort. Conversely, research on if (or when) this decline in T would be followed by an eventual recovery and subsequent shift away from parenting effort is lacking. In a U.S. community sample of 225 males transitioning to first‐time fatherhood, we measured T at three occasions: third trimester, infant 3 months postnatal, and infant 9–10 months postnatal. Using a piecewise latent growth curve model (GCM), we detected a T rebound from when infants were 3 months old to when infants were 9‐10 months old. The slope of this rebound was able to predict paternal care using two distinct measures: (a) an experience sampling method (ESM) that gathered data on paternal time allocation over the course of the study period and (b) independent coders rating fathers for the quality of paternal care during a structured task designed to elicit an infant fear response. As predicted, the more accelerated one’s T rebound (slope), the less time fathers invested in their infants across the study period. However, we found a positive relationship between T rebound and quality of paternal care during a challenging activity. Discussion will focus on nuanced reasons that contribute to these findings as well as speculate on the ultimate function of a human paternal T rebound.