There is already a large literature on car sharing, but understanding its impact on the transportation system needs additional work. One of the main limits of the existing research is that such modes have been mostly analyzed as isolated systems. As they were niche products, it was not easy to include them in comprehensive models of transport, and in a way it even did not make sense. But their current popularity and their expected further growth, completely change the picture. Transport model systems where such mobility concepts are explicitly modeled along all other modes will be crucial in the near future to get an insight on travel behavior impacts and to assess possible future scenarios representing the impact of long term mobility decisions. This paper describes how carsharing demand is modeled within an activity-based multi agent simulation of transport called MATSim (www.matsim.org). It draws from a series of papers written between 2009 and 2015 by the authors and is part of a research effort which ultimate goal is to build a predictive and policy-sensitive model that can be used by practitioners and policy- makers in order to test any type of carsharing scenarios. This paper summarizes the work done, provides some examples of applications, addresses current limitations and shortly reports on ongoing and planned developments.