On April 12 & 13 this year's Annual Meeting and General Assembly of the Confederation of Open Access Repositories (COAR) took place in Vienna/Austria. Through a diverse mix of conference sessions, round-table discussions and open dialogue, the meeting pushed attendees to examine the future of repositories and the challenges and opportunities for international collaboration. Day one had the theme “Next Generation Repositories”. The first presentation was from Herbert Van der Sompel, who built on a recent D-Lib paper to discuss the need for “Establishing New Levels of Interoperability for Web-Based Scholarship”. Van der Sompel argues that we must move from a repository-centric to a web-centric perspective. As he points out, as far as the basic architecture of the web is concerned, there are no such thing as repositories, just individual resources. Hence, repositories need to move beyond reliance on protocols like OAI-PMH, which embodies this viewpoint – addressing interoperability challenges by defining a machine interface for the node and expecting others to use it. Such solutions, says Van der Sompel, merely piggyback on the web without really fully embracing its core technologies. Instead, next-generation repositories should harness the web’s full potential by working with its primary architectural elements (resources, URIs, HTTP, RDF, etc.) to put resources at the centre. This stirring talk presented a call-to-arms to the repository community to reimagine the place and purpose of repositories. In the same session, Pandelis Perakakis discussed how present-day problems with validation, evaluation and dissemination in science all point towards a future of Open Science and how repositories could facilitate this change. Pandelis, a post-doctoral researcher in neuroscience at the University of Granada, has long been an enthusiastic advocate of open science, and is co-founder of Open Scholar, a researcher-led association working towards more open and transparent scientific practices. After an excellent overview of the current problems facing science such as the reproducibility-crisis, the iniquities of the Impact Factor as a metric for scientific excellence and the problems with traditional peer review, Pandelis presented Open Scholar’s two flagship projects: the Self-Journal of Science and the Open Peer Review Module for repositories (the latter initiative the result of an OpenAIRE grant – seehttps://blogs.openaire.eu/?p=775). Wednesday focused on ”International Collaboration” and the question of how COAR can further promote the equitable flow of knowledge across the globe. In the morning Xiaolin Zhang and Li-Ping Ku from the Chinese Academy of Sciences gave an insight into the current state of Open Science in China. Following an overview of the unbelievable growth of the Chinese research and development budget over the last 10-15 years (such that China is now the country with the third most citations globally), Xiaolin Zhang argued the needs for Open Science & Innovation in China, giving an overview of different kinds of infrastructures and outlining future challenges and development. Although China is favourable to Open Access, with national commitments and formal open access policies, nevertheless challenges remain, including the need for stronger commitment for Green OA, the need for further top-down political backing and the need for more responsibility to be taken by libraries. Li-Ping Ku then complemented the view on Open Access in China by focusing on Research Data Management and Open data in China, talking about current policy and government strategy, national efforts, investments and infrastructure as well as future prospects for the coming years. China’s latest Five-Year Plan (2016-2020) includes big data, cloud computing, and ‘internet plus’ as drivers of economic and social development. After lunch came the session on “Bridging Continents”, with perspectives on Open Access from Africa, Asia and North America. First, Daisy Selematsela of South Africa’s National Research Foundation discussed how best to promote exchange and cooperation on OA issues among developing regions. Although the OA agenda is moving forward in South Africa and other African nations, there remains much to be done. Where funding for largescale OA initiatives remains scarce, Selematsela advises that pragmatism is a watchword. For example, by recognising that discovery and access to content/data embargoed in some way is preferable to non-discovery. Next, Kazu Yamaji of the National Institute of Informatics in Japan discussed the OA landscape in his country. Although almost all Japanese national universities now have their own repositories, initiatives like the JAIRO Cloud (a web-hosted repository) is now making repository services available to smaller universities (as of 2015, there were more than 650 repositories in Japan). The next steps lie in furthering interoperability amongst this national network and then furthering that interoperability within Asia and globally (for instance, by investigating data-sharing with OpenAIRE). In the final contribution to the “Bridging Continents” session, SPARC’s Heather Joseph discussed lessons SPARC has learned over the last 10 years of advocating openness. This included invaluable advice such as the need to push the full Open Science agenda (from open data to reproducibility, open source to open peer review), as well as the need to get away from talking about Open Science as an end in itself, but remembering to always discuss it as a means to bettering the way science and education are done. Heather then pivoted to discuss how the changing political landscape in the United States offers a chance to push forward the Open Access/Open Science agenda. The meeting closed with André Laperrière from Global Open Data for Agriculture and Nutrition (GODAN), who discussed the role of GODAN in promoting global open data for agriculture and nutrition. Laperrière gave practical examples of the cross-referencing of open data-sets can help to address a variety of current and future challenges in the agriculture and nutrition sectors. He showed how different stakeholder and interests – economic as well as social, environmental etc. – can be served by opening and connecting data and emphasized the need to make data useful and meaningful, by aligning it with concrete issues and leading cultural changes. Despite several challenges related to data management, licensing, interoperability, sustainability and exploitation, Laperrière left no doubt that open data is necessary to best address a variety of current and future challenges. Good then that COAR and GODAN announced a new partnership which, in conjunction with FAO of United Nations, will address several issues including raising awareness of open access and open science in relation to sustainable development and promotion of policy adoption for data sharing.