Purpose – This paper addresses the question of how clustering effect firms’ capabilities to innovate. In economic geography the traditional view has been that physical proximity within city-regions and being part of an industrial cluster are keys for the innovative capacity of firms. This view is echoed in the management literature which holds a strong belief in the concept of regional clusters. Clusters have been defined as “geographical concentrations of interconnected companies and institutions in a particular field” (Porter 1998). According to Porter, being localized in a cluster improves competitiveness by allowing firms easier access to specialized suppliers and employees; and by increasing firms’ capacity for innovation and productivity growth. However, in a globalizing world the superiority of local clusters should be questioned. Firms that successfully develop global pipelines to knowledge sources in distant locations often sidestep the risk of lock-in, associated with relying solely on local sources of information (Malecki 2010; Fitjar and Rodríguez-Pose 2011). Design/methodology/approach – This paper uses a specifically tailored survey carried out in 2013 of the geographical distance and scope of collaboration for innovation of 2002 firms located in the four largest Norwegian city-regions with the rest of the country as a separate category. Originality/value – We find that a majority of firms are located in regional clusters. However, locating in regional clusters is not the optimal strategy for firms to innovate. Belonging to a national cluster enhances the capability to market new products, and being part of an international business environment improves the capacity to launch radical or substantially new products. The contribution is threefold: Firstly, the comparison of the innovation performance of firms located inside regionally concentrated clusters with those located outside. Secondly, most prior studies, often case studies, have been limited to certain sectors or spatial agglomerations. Using a survey yields greater opportunities than case studies to generalize the findings. And finally, other studies often use patents as a measure of innovation, whereas the present study captures a much broader array of innovations Practical implications – Cluster policies to advance the capabilities to innovate in firms and regions may need to be revised. The policy of initiating and supporting predominantly regional clusters should be substituted, or at least complemented, by a more open and broader spatial approach. Managers may rethink the way their innovation processes are organized, and broaden their search for ideas and knowledge beyond the region they are located.