Purpose The recent economic crisis has significantly slowed Slovenia’s recent social and economic progress and exposed some important long-term problems such as a reliance on low value added industries and lagging labor productivity. The Slovene government has taken steps to create research partnerships between public science and private sector and among multiple private sector companies. In this paper, we conduct a social network analysis of the research partnerships and examine whether public funding has created the desired partnerships. Design/methodology/approach We employed a social network analysis in two stages. In the first stage, we treated the founding partners of government-funded 32 research centers as a single two-mode network and investigated how each of the members was bound to the network. In the second stage of the analysis we used project data from ten of the centers to characterize a project network based on collaborations on specific projects. Thus, the second stage overlaps the center network with the project network. We used information from interviews with network members to assist in interpreting the results. Findings Networking policies are stimulating collaborations among different types of Centers and partners, but to differing degrees. While the formal collaborative network showed strong participation from the private sector, PROs, and HEIs, some of the centers are not well connected to the rest of the network. Partnership in the development of a proposal in response to a tender does not always translate into project collaboration, and the networks have evolved as project workplans and staffing plans are developed. The innovation network is evolving into an international network within and across scientific areas. Networks are path dependent and require policy stability; experienced bridging institutions can fill gaps where partners lack experience. Research limitations/implications Our definition of an a network member is the company, faculty, or department. In reality, individuals within these organizations are acting on their own connections and experiences, and these may or may not encourage other individuals in the same organization to engage in partnerships. Thus, we may be overstating the extent to which one connection among organizations generates experience that will lead to future connection. Another important limitation of our data is that for the second stage of the analysis we received project information from only ten of the 32 formal center programs examined in the first stage. Practical implications Partnership is a learned behavior and the development of trust among partners takes time. The Slovenian government should provide policy stability and allow niches of technical excellence to emerge through consortium proposals. They should monitor the project partnerships and adjust funding so that it is reaching applicants that are actually partnering on projects rather then working alone or within their own institutional types. Other nations should also monitor the impact of partnership programs to ensure that as they evolve the funding is continuing to support and demonstrate the benefits partnership behavior. Originality/value Like many nations struggling to emerge from the recession, Slovenia has had to examine its long-term strategy for upgrading its industries and improving productivity. This paper demonstrates how policies to enhance the innovation agenda might be more effective by examining how the networking resources are actually being used, whether participants are participating in networks that cross institutional types, whether policies are encouraging the exchange of information across stages of the innovation process, and therefore whether the policy will move the culture toward greater collaboration and R&D effectiveness. The results can assist Slovenia’s policymakers in redesigning innovation network policy.