Fossil fuel combustion, land use change and other human activities have increased the atmospheric carbon dioxide (CO2) abundance by about 50% since the beginning of the industrial age. The atmospheric CO2 growth rates would have been much larger if natural sinks in the land biosphere and ocean had not removed over half of this anthropogenic CO2. As these CO2 emissions grew, uptake by the ocean increased in response to increases in atmospheric CO2 partial pressure (pCO2). On land, gross primary production also increased, but the dynamics of other key aspects of the land carbon cycle varied regionally. Over the past three decades, CO2 uptake by intact tropical humid forests declined, but these changes are offset by increased uptake across mid‐ and high‐latitudes. While there have been substantial improvements in our ability to study the carbon cycle, measurement and modeling gaps still limit our understanding of the processes driving its evolution. Continued ship‐based observations combined with expanded deployments of autonomous platforms are needed to quantify ocean‐atmosphere fluxes and interior ocean carbon storage on policy‐relevant spatial and temporal scales. There is also an urgent need for more comprehensive measurements of stocks, fluxes and atmospheric CO2 in humid tropical forests and across the Arctic and boreal regions, which are experiencing rapid change. Here, we review our understanding of the atmosphere, ocean, and land carbon cycles and their interactions, identify emerging measurement and modeling capabilities and gaps and the need for a sustainable, operational framework to ensure a scientific basis for carbon management.