The automotive industry has been facing a great challenge on the current market. With the rapid advance of technology and the growth complexity of its projects, one question comes up: how to speed up development time, while increasing options for the products, maintaining the quality and keeping the competitiveness at the same time? The answer for this question is not easy and can include several decisions and actions in the whole development process. In this context, partial system virtualization, or even a complete virtualization of systems included in a vehicle, can be used as an important tool to help answer the mentioned question. In other words, the usage of modeled components can help decrease the development time and increase product quality. Besides, this virtualization can be used on different steps during the development of new products or even for the maintenance of those already on the market. There are several decisions associated with this virtualization that needs to be taken into account in order to balance the cost benefit ratio, such as: 1) which development phase the models will be used; 2) how accurate the models needs to be; 3) will it use only models, only real components or a mix between them; In other words, the requirements of these tools are intrinsically related with the development phase, such as concept, prototyping, test, validation and so on. In this context, this article will discuss aspects that needs to be considered when the requirements of automotive systems virtualization are defined. Some examples will be presented to illustrate how these models can be used and how they can help in the automotive development process.