Scientific advancement depends on sustaining a workforce trained in the scientific method. These scientists can in turn train the next generation of scientists and ensure that novel research continues to contribute to society. However, recent evidence shows that academic careers are shortening over time (Milojević et al., 2018). This coupled with a surplus of early career researchers (ECRs) relative to academic faculty positions raises the concern that talented scientists might not be retained in academic research. However, the empirical properties of the modern academic workforce are largely unknown. We use the publication histories of 3.5 million researchers to examine how career age composition of the scientific workforce has changed over 30 years. We find that for most fields there has been a relatively stable distribution of scientists at various career stages, with notable exceptions in Health and Physics. Using these data, we calculate the rate that researchers enter and depart the academic workforce, finding the most rapid turnover among ECRs. While varying the rates of turnover and replacement based on these observations, we develop a model and project a continuation of current career age proportions. This implies that the makeup of the workforce may remain stable with an abundance of ECRs. We conclude that the disconnect between the number of ECRs and the number of available long-term academic positions may be attributed to the collaborative nature of the academy. The stabilization of the workforce with a high proportion of ECRs urges a reassessment of how academic careers are portrayed.