Materials science is greatly benefiting from computational data being openly accessible through various materials encyclopedias. Also, open access peer-reviewed publishing has become the norm rather than the exception. In contrast, the publication of experimental research data in materials science still is in its infancy. I will discuss the requirements for publishing experimental research data according to the FAIR (Findable, Accessible, Interoperable, Reusable) principles. Given the size and complexity of data generated in decentralized lab-based experiments, FAIR data publishing requires open data formats, community-wide metadata conventions, and transparent and reproducible data processing workflows.