The chapter explores the concept of data portability in a data-driven world. In the first part, it maps out the journey data takes in a data economy and investigates the valuation and cost of data. It posits that, because of data analytics, machine learning, and artificial intelligence models, “generated” data, as data that has been derived or inferred from “raw” data, is of higher value in the data market, and carries a higher cost of production. In the second part, building on this taxonomy, the requirements for the free flow of data in competitive data-centric markets are discussed: regulations on data tradability and portability. The analysis leads to doubt that the newly introduced and widely debated rules regarding portability of data under European Union law will adequately provide these prerequisites. The chapter concludes by suggesting an alternative model for data portability: rather than distinguishing between personal and non-personal data, the legal concept of data portability should be based on the allocation of value; that is, whether the value of data provides cost compensation to the service provider. Raw data should always be portable, while generated data may require a more refined regime depending on whether it is compensating costs.