Previous attempts to quantify the amount of sedimentary carbonate precipitation in modern marine sediments have been derived from the flux of calcium into the sediments due to diffusion under assumed steady state, the application of Fick's first law, and then extrapolation of these site-specific rates to the global ocean sediment column. This approach is limited, however, as much of the ocean floor has not been sampled. We take a machine learning approach to update and refine the estimate of the amount of sedimentary carbonate precipitation, as well as define whether sedimentary carbonate precipitation is driven by organoclastic microbial sulfate reduction or anaerobic methane oxidation. We identify areas where there is sedimentary carbonate formation using machine learning, based upon oceanic physical and chemical properties including bathymetry, temperature, water depth, distance from shore, and tracers of primary production, and data from the global ODP/IODP database. Our results suggest that the total amount of sedimentary carbonate formation is much lower than previous estimates, at 1.35±0.5x1011 mol C/yr. We suggest that this rate is a lower estimate and discuss why machine-learning approaches may always produce lower-bound estimates of global processes. Our calculations suggest that the formation of sedimentary carbonate today is mainly driven by anaerobic methane oxidation (77%), with the remainder attributed to organoclastic sulfate reduction. We use our machine-learning results to speculate the impact that sedimentary carbonate precipitation may have had on the carbon isotope composition of the surface dissolved inorganic carbon reservoir over Earth history.