An exhaustive review was undertaken to assemble all available correlations for supercritical CO2 in straight, round tubes of any orientation, with special attention paid to how the wildly varying fluid properties near the critical point are handled. The assemblage of correlations, along with subsequent discussion, is presented from a historical perspective, starting from pioneering work on the topic in the 1950s to the modern day. Despite the growing sophistication of sCO2 heat transfer correlations, modern correlations are still only generally applicable over a relatively small range of operating conditions, and there has not been a substantial increase in predictive capabilities. Recently, researchers have turned to machine learning as a tool for next-generation heat transfer prediction. An overview of the state-of-the-art predicting sCO2 heat transfer using machine learning methods, such as artificial neural networks, is also presented.