The project aims in fact at defining a complete toolchain that, starting from models of CPS basic components, functions and prototype behaviors, enables the designer to set-up collaborative autonomous CPSs, to test the swarm performance with respect to the design goal (i.e., to evaluate the solution fitness against the design requirements) and to massively deploy solutions to “reconfigurable” CPS devices. CPSwarm builds upon well-known initiatives, and state-of-the-art, in CPS design, such as the Ptolemy project and the Action Webs research , and it considers autonomic and swarm computing as thrusters to address the enormous potential, and risks, represented by billions of networked sensing and actuating devices deployed worldwide. Model-centric design is the spine of the CSwarm project, which, in fact, aims at providing a library of easy to reuse models for CPS design. Additionally, predictive engineering is another pillar of the project, enabling model verification and simulation of collaborative, autonomous, CPS behavior against real-world and hard-to-handle physical data. Such an approach pushes forward CPS engineering at a larger scale, with expected significant reduction of development time and total cost of ownership.