Estimates of biogeochemical properties should ideally be reported with uncertainty. But what are the consequences of that uncertainty for real world decisions, applications and future research? We recently published the world’s first fully data-constrained global terrestrial carbon model – in which all parameters of a simple process-based carbon model have been inferred as probability distributions from empirical datasets on carbon stocks and fluxes. This estimates potential terrestrial carbon storage for every locality on earth as a probability distribution. Here we explore the implications of that uncertainty for Agriculture, Forestry, and Land Use change (AFOLU) projects aiming to generate money from carbon fixation and storage. We estimate that, at $20 per ton avoided CO2 emissions, further reducing uncertainty in the model parameters alone would generate thousands of additional dollars per hectare for individual projects, exceeding returns from crops and timber in many places, and of the order of billions of additional dollars for global carbon markets overall. This shows a very real financial incentive for performing research to further reduce uncertainty in terrestrial carbon estimates as well as a financial measure of the impact of performing additional research.