Future cloud-based, service-oriented training environments, such as the Army's planned Synthetic Training Environment (STE), are expected to consume authoritative data sources for terrain and simulation models, and to stimulate and consume data from mission command systems. This paper describes the envisioned capabilities of the STE in general, and the training management tools in particular. STE is anticipated to be supported by training management tools that will use unit training records and plans to help automatically or semi-automatically tailor STE training exercises to the unit's current training needs. At issue is how these adaptive training management services will exchange data with other STE components. The paper argues that semantic interoperability will be required. A review of existing and developing data exchange standards in the modeling and simulation domain and in the adaptive training technology domain suggests they are unlikely to support the semantic interoperability required. It is suggested that the National Information Exchange Model (NIEM) may represent a possible method of providing that level of interoperability. NIEM, which is federally governed, allows communities of interest to establish a common vocabulary and to use it to create standardized machine-readable information exchange packages. These data packages use World Wide Web Consortium (W3C) Extensible Markup Language (XML) schema or NEIM-Unified Modeling Language. In 2013 the Chief Information Officer for the Department of Defense (DoD) issued the “NIEM firstâ€� memorandum, which directed that the DoD shall consider NIEM first for their data exchange standards. A MilOps community of interest has used NIEM successfully to exchange information with coalition partners. Of additional relevance, work in the geospatial community has demonstrated the combined use of NIEM, intelligence community security specifications, Open Geospatial Consortium web services and Geography Markup Languageaware clients to support information exchange among authorized users. The paper recommends that STE proponents consider reuse of and building upon prior NIEM work to support the semantic interoperability required among the models, simulations, authoritative data sources, and training management systems that will make up the STE.