Determination of agricultural fields carbon budget is needed to devise the best sustainable agriculture strategies, and quantify future subsidies in the agricultural sector. One major sustainable practice is cover crops that are planted in between main crops cycles, enabling storage of carbon and reduction of reflected heat. Assimilation of Earth Observation data is crucial for such exercise due to the sheer heterogeneity and complexity of crop modeling in the real environment. In this paper, we present the impact of cover crops on Net Ecosystem Exchange (NEE) and Biomass over a large area in the south-west of France using the AgriCarbon-EO. This tool is an end-to-end solution that enables a Bayesian assimilation of multi-temporal and entire tiles (100x100 km) Sentinel-2 reflectance data into a radiative model and a crop model. Our results show an increase of NEE of about due to cover crops over maize fields.