Earth system models (ESMs) highly simplify their representation of biological processes, leading to major uncertainty in the impacts of climate change. Despite a growing understanding of molecular networks from genomic data, describing how changing phytoplankton physiology affects biogeochemical processes remains elusive. Here, we embed genome-scale models within a state-of-the-art ESM to deliver an integrated understanding of how gradients of nutrients modulate the molecular physiology of various plankton. In particular, when applied to , we find that glycogen and lipid management can be interpreted in terms of acclimation to different environments. Generalized to other phytoplankton such as the diatom , we estimate the production of 39 metabolites that constitute hot spots of dissolved organic carbon described by their amount of carbon produced and their diversity of associated metabolites in ESMs. This modeling approach shows how genome scale-enabled ESMs have the potential to advance our understanding of microbial ecosystem functioning in ocean biogeochemical processes.