The Data Mesh paradigm has emerged as a promising response to the limitations of centralized data architectures by decentralizing ownership and promoting the principle of “Data as a Product”. Although promising, its implementation hinges on the management of distributed data assets, particularly in terms of governance and interoperability. This paper investigates how the Findable, Accessible, Interoperable, Reusable (FAIR) data principles can address these gaps and strengthen the foundations of Data Mesh. Through a systematic literature review and the design science research methodology, we investigate the relation between the FAIR and the Data Mesh principles, revealing a strong conceptual alignment. Based on these findings, we extend a Data Mesh Reference Architecture using the ArchiMate enterprise modeling language, providing a blueprint that incorporates the FAIR data principles into organizational processes, data product lifecycles, and governance mechanisms. Our contributions are threefold: (1) a systematic mapping of FAIR principles to Data Mesh principles, (2) an extended reference architecture that demonstrates their integration in practice, and (3) concrete technical recommendations, such as the use of persistent identifiers and JSON-based semantic schema, to implement FAIR-compliant Data Products. Validation is done through schema testing, and a music streaming platform case illustrates the feasibility and benefits of the approach. This study offers to both scholars and practitioners actionable models and guidelines for designing FAIR Data Mesh architectures, paving the way for empirical research on real-world implementations.