Adaptive learning means offering customized educational environments for each individual. For this to be possible in virtual environments, like the Internet, actual standardized data models for e-learning must be enriched with explicit “machine-processable” semantic information and appropriate inference rules. This would allow Learning Management Systems (LMS) to automatically process and integrate these models meaningfully in order to offer alternative behaviours depending on each learner and his particular needs. This paper analyses the impact of the upcoming Semantic Web, an extension of the current Web in which information is given an explicit meaning, on the consecution of real high-quality adaptive e-learning environments.