Yihan Wang, Sébastien Bourdin, Fabien Nadou, Emmanuelle Faure (p), Romain Gibert

Michael Fritsch

Innovation intermediaries are the public or private organizations that support innovation activities by facilitating knowledge sharing and collaboration among various stakeholders in a regional innovation ecosystem. They play the crucial roles in the territorial governance of regional production and innovation activities, whereby different parties or actors of various natures contribute to working out—sometimes through discussion, and sometimes through conflict—common projects for the future development of the territories. By coordinating innovation projects and organizing social events, innovation intermediaries stimulate technological and market knowledge diffusion and strengthen the ties among stakeholders of regional innovation ecosystem.

Meanwhile, the geographical proximity of stakeholders and the organized proximity in innovation cooperation affect the business model of innovation intermediaries, in turn, they play different roles in territorial governance mechanism. Nonetheless, limited research has been done on how different types of innovation intermediaries affect regional innovation in the territorial mechanism based on geographical and organized proximity of stakeholders.

This research studies how different modes of innovation intermediaries affect territorial governance of regional innovation. Specifically, we focus on two types of public innovation intermediaries in France - "pôles de compétitivité" (pôles) and "filières d’éxcellence"(filières). Whereas they both gather multiple stakeholders in the regional innovation ecosystem that collaborate on common interests, they play different roles in the territorial governance of regional innovation. The “filières” are both more territorially rooted and more focused on connecting local stakeholders on specific sectors. In contrast, the “pôles” establish cross-sectorial collaborations among stakeholders in different regions.

In this research, we compared the modes and roles of "pôles" and "filières" in the territorial governance of regional innovation in Normandy, France. Specifically, we focus on the geographical location of affiliates, the organizational logic and their functions in intermediation of both types of public innovation intermediaries. Based on a survey of members from both mechanisms, as well as interviews with local actors, this study will define the specific characteristics of each territorial governance model. Thereafter, we compare how these two types of innovation intermediaries balance cooperative and competitive dynamics in regional innovation and affect the network structure of partnerships

These findings will contribute to the literature on the territorial governance of innovation intermediaries and provide insights for policymakers responsible for their management and coordination, which can present challenges within the framework of Regional Innovation Systems.

Michael Fritsch (p), Ann-Marie Sommerfeld, Michael Wyrwich

Alina Sorgner

We investigate regional innovation output in terms of patenting per population over the period 2000-2020 in the European NUTS2 regions.1 While innovation activities in European regions exhibit in general a pronounced tendency of convergence, some of the lagging regions show at best only little improvement and fall further behind. One might say that these regions are captured in an innovation trap, a trajectory that constitutes a certain type of development trap. Development traps may indicate a dynamic ‘market failure’ that suggests a case for policy intervention.

In a first step, we discuss several approaches to identify regions that may be considered to be in an innovation trap. We then analyze the characteristics of the regions that are considered problematic according to the different definitions. Comparing them to regions with low levels of innovation activity in the base year but significant improvements over time, aim at identifying reasons for stagnation versus drivers of growth. In the analysis, we also regard different measures for the types and quality of patents such as ‘radicalness’ and ‘originality’. Overall, regions that may be considered to be in an innovation trap are a rather heterogeneous group. This implies that there is no one size fits all approach when it comes to policies aimed at improving innovation performance but rather specific regional characteristics should be taken into account.

Alina Sorgner (p)

Tingting Liu

This paper investigates the regional distribution of firm-level AI technology adoption using a novel source of web-scraped data from websites of economically active firms in Europe. Using a narrow definition of AI adoption based on the importance of AI for a firm’s organizational identity, it is shown that AI adoption remains a very rare phenomenon in Europe occurring on average in 0.26 percent of all firms. Moreover, about half of all AI adopting firms are young startups that are no older than five years. While AI adopters are often based in highly dense urban areas, often in capital city metropolitan regions, regions with a high share of AI startups among all AI adopters are not necessarily big cities. Regional factors that are associated with the firm-level AI adoption are, for instance, regional demographics (e.g., lower median population age), regional socioeconomic characteristics (e.g., higher GDP per capita, higher total factor productivity (TFP)), regional digital economy (e.g., a higher share of employment in high-tech sectors and digital users), digital and physical infrastructure. Regional factors seem to play a less pronounced role for AI startups than for established AI adopters. Interestingly, the share of AI startups in all AI adopters is higher in regions with a significantly lower TFP. These results point toward emerging AI entrepreneurship ecosystems in Europe, for which geography plays a potentially different role compared to traditional entrepreneurship.

Tingting Liu (p), Deyu Li, Kerstin J. Schaefer

Carolina Foglia

As Big Tech firms such as Google expand their global reach, their internationalization strategies have attracted increasing attention. However, current studies focus mainly on the business expansion of Big Tech firms, and little attention has been paid to their R&D locations. Besides the well-acknowledged region-specific factors, Big Tech firms’ R&D is affected by technology-specific factors as well, as their inventions show significant technological heterogeneity. Specifically, Big Tech firms develop both core technologies, characterized by high levels of technological continuity, and application technologies that are beyond enhancing ICTs and are more recombinative. Such technology distinction requires various knowledge bases, yet the underlying mechanisms remain unpacked. Based on patent data in 2010-2020, this study investigates the location choice of Big Tech firms’ R&D internationalization, examining the role of knowledge relatedness, i.e., similarity and complementarity, between host locations and Big Tech firms. We expect to find an overall positive association between Big Tech firms’ inventions and knowledge relatedness with host regions. Furthermore, Big Tech firms are more likely to locate their R&D in terms of core technologies (core RDI) in regions with a similar knowledge base to ICT fields while locating those related to application technologies (apply RDI) in regions with a complementary knowledge base. This study links evolutionary economic geography to the international business studies, highlights the distinct role of knowledge relatedness, and provides insights for regional innovation policy-making in the digital era.

Roberta Capello, Carolina Foglia (p), Camilla Lenzi

Emmanuelle Faure

The fast development of Artificial Intelligence (AI) in recent years and the widening of AI applications in a variety of domains have raised concerns about its impact on market restructuring and the concentration of power in the market and across space.

Based on a large-scale matched patent-firm dataset sourced from Bureau Van Dijk product ORBIS-IP including AI patents filed at the European Patent Office between 1977 and 2020, identified through text mining techniques, this paper reconsiders this statement by proposing a conceptual distinction between entities participating in the invention of AI technologies and those involved in its development and commercialisation.
Based on this distinction, the paper aims to offer a taxonomy of AI invention and development at the regional level, and to highlight whether (and which types of) regions previously at the margins of the current technological landscape could approach the AI technological frontier.

This exploration sheds light on how niche realities can scale up and become significant players in the AI era, thus offering new reflections for policy fostering regional innovation and transformation.