The System Architecture Virtual Integration (SAVI) program is a collaboration of industry, government, and academic organizations within the Aerospace Vehicle System Institute (AVSI) with the goal of structuring a new integration process that relies on a single-truth architectural framework. The SAVI approach of Integrate, then Build provides a modern distributed development environment which arrests the propagation of requirements errors through the development life cycle. It does so by capturing design assumptions and shared properties of the system design in an authoritative, annotated architectural model. This reference model provides a common, analyzable framework for confirming that system requirements remain complete, consistent, and correct at all levels of system decomposition. Core concepts of SAVI include extensive use of model-based system engineering tools and use of a single-truth reference architectural model. From the outset SAVI developers anticipated that a quantified prediction of the productivity of the SAVI Virtual Integration Process (VIP) would be necessary to close the business case for using it. Therefore, the SAVI statement of work at each stage of its feasibility demonstration carried a task to estimate the Return on Investment (RoI). AVSI participants needed a prediction of what the resources poured into SAVI development and deployment would produce. This paper lays out the work done so far to produce these RoI estimates and the assumptions that have gone into them. The paper goes on to illustrate example results for two of the major types of participants in SAVI, and details the current state of the evolving estimation capability. The most important result of this RoI work is a substantial positive RoI predicted for using SAVI's VIP. Initial estimates of RoI for a first application to a commercial aircraft development indicated an expected value of annual RoI for an OEM on the order of 40%. Later estimates gave similar, but more positive, results with modified assumptions. But the range of variation of the estimates has been reduced to less than 1/3 of the original prototype estimator's variation. Savings for suppliers heavily engaged and at risk in the development are also predicted to have double digit annual RoIs, with the exact value of annual RoI rate dependent on the level of involvement of the supplier. The minimum value of annual RoI for the same commercial airliner development was calculated to be 2%, using the initial prototype estimator, but that minimum value of annual arithmetic RoI grew to over 70% per year in the refined estimator.