Since the 2000s, European open data policies have given a strong boost to commercial meteorology by giving free access to weather observations and models produced by public organizations. This article examines the efforts and challenges met by a French company that developed an offer of weather services based on the commodification of both open weather data and local observations produced by low-cost stations used by farmers. However, the paper shows that such commercialization of stations’ data is hampered both by their (...) material friction with the weather infrastructure and by an economic friction related to data’s heterogeneous valuations. As solving the economic friction generated by data’s valuation is the priority of commercial meteorologists, they finally differentiate their offer using simulated weather data produced by computing models. Finally, the paper claims that the commercial process inherent to the use of this other type of data subordinates weather science to a logic of services: when designing a service for agriculture, the company considered that simulated data were more suited to digital farm advising services. In the end, it is the capacity of weather data to be integrated into private services and tools that valuate them as suited to guide farming practices. (shrink)