Navigating Hydrogen Horizons - Mapping the Technological Innovation System of Mid Sweden Hydrogen Valley A case study in collaboration with CIT Renergy Master’s thesis in Management and Economics of Innovation TOM BANJAC MELVIN SÖRUM DEPARTMENT OF TECHNOLOGY MANAGEMENT AND ECONOMICS DIVISION OF INNOVATION AND R&D MANAGEMENT CHALMERS UNIVERSITY OF TECHNOLOGY Gothenburg, Sweden 2024 www.chalmers.se Navigating Hydrogen Horizons - Mapping the Technological Innovation System of Mid Sweden Hydrogen Valley A case study in collaboration with CIT Renergy TOM BANJAC MELVIN SÖRUM Department of Technology Management and Economics Division of Innovation and R&D Management CHALMERS UNIVERSITY OF TECHNOLOGY Gothenburg, Sweden 2024 Navigating Hydrogen Horizons - Mapping the Technological Innovation System of Mid Sweden Hydrogen Valley A case study in collaboration with CIT Renergy TOM BANJAC, 2024 MELVIN SÖRUM, 2024 © TOM BANJAC, 2024 © MELVIN SÖRUM, 2024 Department of Technology Management and Economics Chalmers University of Technology SE-412 96 Gothenburg Sweden Telephone +46 (0)31-772 1000 Acknowledgements, dedications, and similar personal statements in this thesis, reflect the authors‘ own views. Navigating Hydrogen Horizons - Mapping the Technological Innovation System of Mid Sweden Hydrogen Valley TOM BANJAC & MELVIN SÖRUM Department of Technology Management and Economics Chalmers University of Technology Abstract The transition towards a sustainable energy future necessitates the adoption of low-carbon technologies, with hydrogen emerging as a pivotal element in achieving climate neutrality. Hydrogen valleys – a geographical area where several hydrogen applications are combined into a more cost-efficient and integrated hydrogen ecosystem, are identified by the European Commission as important enablers for transitioning into a full-scale hydrogen economy. This master’s thesis investigates Mid Sweden Hydrogen Valley (MSHV) – a hydrogen valley in early stages of formation in Sweden, using a Technological Innovation System (TIS) framework. Due to the novelty of hydrogen valleys, few studies have analyzed hydrogen valleys using system level frameworks like TIS – which could provide valuable insights regarding how a hydrogen valley should be developed. By employing a qualitative case study approach, data was collected through 21 interviews with members of MSHV and actors closely related to it. Key findings reveal that a hydrogen valley is a complex system, including diverse networks of actors both in the main hydrogen value chain, through supporting value chains, and other influences such as grant providers and authorities. The findings also reveal that MSHV performs relatively well based on the appraisal of its seven functions. Still, six challenges (labeled blocking mechanisms) are identified that prevent the hydrogen valley from moving smoothly from a formative phase into a growth phase. These challenges are addressed by proposing six recommendations for policymakers and three recommendations for managers, which emphasize the need for collaboration, improved economics, and political support. Finally, the study concludes by suggesting directions for future research. Keywords: Hydrogen Valley, Technological Innovation System, Fossil-free Hydrogen, Low-carbon Hydrogen, Socio-technical Transformation, Sustainable Transformation Acknowledgements This master’s thesis was conducted in collaboration with CIT Renergy and completed at the Department of Technological Management and Economics at Chalmers University. Firstly, we extend our deepest thanks to Viktor Stenberg and Pontus Bokinge at CIT Renergy, for their support and guidance throughout this entire journey. Their valuable contribution and continous encouragement have significantly contributed to the success of our work. Secondly, we would like to express our sincere gratitude to our supervisor Linus Thomson and examiner Ksenia Onufrey at Chalmers, for their dedicated time, effort and insightful feedback during this spring. Linus’ consistent support and continous valuable feedback has been crucial in advancing our research, from day one. Ksenia has also provided insightful feedback during several sessions which has greatly enhanced the quality of our work. Lastly, we appreciate all the interviewees that partcipated in our study. Their valuable insights and perspectives formed the foundation of our data collection and were essential in realizing this thesis. We are deeply grateful for their contribution. Tom Banjac & Melvin Sörum Table of contents Abstract ....................................................................................................................................................................... 6 Acknowledgements .................................................................................................................................................. 8 1. Introduction ......................................................................................................................................................... 13 1.1 Background ............................................................................................................................................................................................ 13 1.2 Aim .............................................................................................................................................................................................................. 17 1.3 Delimitations ......................................................................................................................................................................................... 17 1.4 Specification of the issue being investigated ......................................................................................................................... 17 2. Theory ................................................................................................................................................................... 18 2.1 Introduction to technological innovation systems ............................................................................................................. 18 2.2 Constituents of technological innovation systems .............................................................................................................. 19 2.2.1 Structural components ........................................................................................................................................ 19 2.2.2 Functional analysis ............................................................................................................................................... 20 2.2.3 Inducement and blocking mechanisms .......................................................................................................... 24 2.3 Critique ..................................................................................................................................................................................................... 25 3. Method................................................................................................................................................................... 27 3.1 Framing research ................................................................................................................................................................................ 28 3.2 Data collection ...................................................................................................................................................................................... 29 3.2.1 Interviews ................................................................................................................................................................ 29 3.2.2 Secondary data collection ................................................................................................................................... 32 3.3 Analysis ..................................................................................................................................................................................................... 32 3.3.1 Thematic analysis .................................................................................................................................................. 33 3.3.2 Steps in the analysis of technological innovation systems ...................................................................... 33 3.3.3 Our analytical framework ................................................................................................................................... 37 3.4 Reflection on validity ......................................................................................................................................................................... 38 4. Analysis ................................................................................................................................................................. 39 4.1 Structural mapping ............................................................................................................................................................................ 39 4.1.1 Actors......................................................................................................................................................................... 40 4.1.2 Networks .................................................................................................................................................................. 45 4.1.3 Institutions .............................................................................................................................................................. 49 4.1.4 Infrastructure ......................................................................................................................................................... 51 4.1.5 Further structural considerations ................................................................................................................... 52 4.2 Functional analysis ............................................................................................................................................................................. 54 4.2.1 Knowledge development and diffusion ......................................................................................................... 55 4.2.2 Influence on the direction of search................................................................................................................ 58 4.2.3. Entrepreneurial experimentation ................................................................................................................... 60 4.2.4 Market formation .................................................................................................................................................. 62 4.2.5 Legitimation ............................................................................................................................................................ 64 4.2.6 Resource mobilization ......................................................................................................................................... 65 4.2.7 Development of positive externalities............................................................................................................ 66 4.3 Identifying inducement and blocking mechanisms............................................................................................................ 67 4.4 Specifying key policy and managerial recommendations .............................................................................................. 72 5. Conclusion ............................................................................................................................................................ 79 6. References ............................................................................................................................................................ 82 13 1. Introduction In this section, an introduction is given to the study, which includes: (1) background, (2) aim, (3) limitations, and (4) research questions. The thesis is thereafter structured by the following: a description of the theory that is used throughout the report, a method chapter where the research process is presented, an analysis of the collected data using an analytical framework, and finally, the research questions are answered in the conclusion, followed by limitations as well as suggestions for further research. 1.1 Background Achieving the future climate goals outlined by the European Union, including climate neutrality by 2050 (European Council, 2022), requires decarbonization of various sectors. Many studies agree that hydrogen will be critical in reaching the sustainability goals (Jodry et al., 2023; Majka et al., 2023; Weichenhain et al., 2022; Ficco et al., 2022; López et al., 2023). It is estimated that hydrogen will account for more than 13% of the world's total energy consumption by 2050, whereas this number was 1% in 2020 (López et al., 2023). This opportunity for growth in hydrogen demand can be attributed to its characteristics, for example: close to zero pollution when integrated with renewable energy sources, flexibility and sector coupling possibilities (i.e. interconnecting the energy consuming sector with the energy producing sectors), and use-cases in several sectors (Genovese et al., 2024). Further, studies such as Ficco et al. (2022) and López et al. (2023) explain that hydrogen could be used to neutralize the emissions of the so called “hard-to-abate” sectors where electrification is currently challenging, e.g. long-distance and heavy goods transport. Other industries that could benefit greatly from hydrogen include the chemical and steel industries, primarily because hydrogen does not emit any CO2 when consumed unlike other fuels (Ficco et al., 2022). Today, most hydrogen is so-called gray hydrogen, meaning that it is derived from fossil fuels – for example through reforming of natural gas. The total worldwide hydrogen production, which was 94 megatons in 2021, resulted in 900 megatons of CO2 emissions. Only 0.7% was produced by low- emission methods, with even less (35 kilotons) being produced through water electrolysis (López et al., 2023). In Sweden, there are currently no official statistics on hydrogen usage, but Fossilfritt Sverige (2021) estimated that the country’s total hydrogen production and usage was 180 thousand tonnes per year (approximately 6TWh/year). The hydrogen was mainly fossil-based, with less than 3% being produced through water electrolysis. According to the report, most of the hydrogen produced in Sweden is consumed near to its production sites. 14 To reach the sustainability goals, it is estimated that all hydrogen production will need to be either green hydrogen (i.e. water electrolysis powered by renewable energies such as wind or solar) or blue hydrogen (i.e. using the same methods as gray hydrogen but with carbon capture and storage) by 2050 (López et al., 2023). Further, López et al. (2023) highlights how different agencies have come up with scenarios for the future hydrogen demand, which are summarized in Figure 1. As shown in the projections, the current production of hydrogen by low-emission methods is practically negligible compared to what is needed in the future. There also exist other forms of hydrogen which are not mentioned as frequently in the literature, e.g. pink hydrogen, which is produced through nuclear powered electrolysis (Shirizadeh & Quirion, 2023). Throughout this study, the term hydrogen will be used to refer to low-carbon hydrogen (unless stated otherwise), which includes all production methods with a carbon footprint close to zero. Figure 1. Projections of future hydrogen demand (López et al., 2023). The European Union sees low-carbon hydrogen as a decarbonization enabler, pushing different policies to support its development. For example, the “REPower EU” plan aims to speed and scale up usage and production of low-carbon hydrogen and to double the amount of hydrogen valleys in Europe to reach 100 by 2030 (Weichenhain et al., 2022). A hydrogen valley is defined by the European Commission as: “a geographical area – a city, a region, an island or an industrial cluster – where several hydrogen applications are combined together into an integrated hydrogen ecosystem that 15 consumes a significant amount of hydrogen, improving the economics behind the project. It should ideally cover the entire hydrogen value chain: production, storage, distribution and final use.” - European Commission (n.d.) Even though hydrogen valley formation is in its infancy, it is expected to be a considerable driver towards a full-scale hydrogen economy (Weichenhain et al., 2022; López et al., 2023). By leveraging synergies through collaboration between actors, hydrogen valleys have the possibility to overcome many of the challenges associated with novel and disruptive technologies. Such challenges include deficient cost competitiveness, lack of scale in production, and infrastructure needs (International Energy Agency, 2019; López et al., 2023). For example, the production of blue hydrogen, which requires CO2 capture, is currently an expensive process, costing $1.4-$4.7 per kilo to produce, while green hydrogen is even more costly at $4.5-$12 per kilo – compared with gray hydrogen which only costs $0.98-$2.93 per kilo (BloombergNEF, 2023). Another example is the need for infrastructure, e.g. pipelines, storage facilities and equipment for loading and unloading, that would allow for distribution of large volumes of hydrogen from producer to end users (Steen, 2016) – which currently does not exist. These challenges constitute important reasons why collaborating, producing, and using hydrogen within a confined geographical area, like a hydrogen valley, is beneficial. A hydrogen valley under formation in Sweden – called Mid-Sweden Hydrogen Valley (MSHV), will be the case of analysis. The mid-Sweden area has a well-established industry, with strong competence within hydrogen, making it an appropriate geographical area to create a hydrogen valley (Region Gävleborg, 2022). Recent developments in the area include the University of Gävle taking the lead of hydrogen research, efforts by Gävle Harbor to establish an energy-optimized harbor cluster (Region Gävleborg, 2022), Inlandsbanan planning for large-scale hydrogen distribution and converting diesel- powered freight trains to hydrogen power (Inlandsbanan, 2021), hauliers like MaserFrakt launching the first hydrogen-powered heavy trucks on the roads in Sweden (Hynion, 2024), and steel companies like Ovako which invest in large-scale electrolyzers for hydrogen production (Vätgas Sverige, 2023). The actor roles in MSVH’s value chain include renewable energy providers, hydrogen producers, hydrogen storage/distribution providers, hydrogen users, grant providers, incubators and research institutes, authorities, and OEMs. MSHV was formed in 2021 based on the shared interest of authorities and industry to strengthen the regional industry, enhance collaboration, and to reach the climate goals (Region Gävleborg, 2022). As it stands, the organization is not a legal entity, which means that MSHV itself cannot initiate projects or allocate funds. The organization is instead used as a platform for collaboration and knowledge-sharing between the members. Coordinated by the regional government, it is free to join for any actor involved in or aspiring to be part of the hydrogen value chain. The hydrogen valley is still in its infancy and does not fully align with the European Commission's definition yet, as many projects are still in the ideation phase rather than the implementation phase. 16 Because the development of hydrogen valleys is a novel phenomenon, both in theory and practice, it raises a series of questions. Potential questions with hydrogen valleys relate to (1) how the hydrogen valleys should be structured, (2) what influences its development, and (3) what challenges and opportunities exist that both policymakers and managers need to address to create a well-performing hydrogen valley. Given the pressing need to scale up low-carbon hydrogen production and given that there are currently many challenges that prevent it, research needs to be conducted to support managers and policymakers when navigating this challenging terrain. In the innovation systems literature, much has already been written about how policymakers should act in similar environments, e.g. the development of technological innovation systems (TIS) for novel energy solutions (see Bergek et al., 2008a), but none have analyzed the specific case of hydrogen valleys. The literature on technological innovation systems is used for the analysis in this thesis. TIS can be defined as a “network of agents interacting in the economic/industrial area under a particular institutional infrastructure and involved in the generation, diffusion and utilization of technology” (Carlsson and Stankiewicz, 1991). This provides a promising framework for analyzing hydrogen valleys due to several reasons. Firstly, it provides a systemic perspective considering actors, networks and institutions, and the interactions between them (Bergek et al., 2008a). Hydrogen valleys – a socio- technical system aimed at contributing to a sustainable future – involve many stakeholders, ranging from companies within the value chain, supporting actors like research institutes and incubators, to governments. The TIS framework could therefore be used to address the first question by understanding the interactions between these stakeholders and gaining insights about how the hydrogen valley should be structured. Secondly, the TIS framework allows the researcher to analyze the dynamics of some key processes (called functions) that influence the development, diffusion and use of a new technology (Bergek et al., 2008a). Exploring how different functions in the TIS are affecting the development of a hydrogen value chain, and linking these to the structural components, could reveal insights about the strengths and weaknesses of the TIS. The functions could therefore prove useful when addressing the second question regarding what factors influence the hydrogen valley development. Lastly, building on the previous two reasons, TIS offers a holistic understanding of innovation dynamics, including technological, legal and socio-economic perspectives (Bergek et al., 2008a), which therefore provides an integrated and systemic approach for studying hydrogen valleys. Accordingly, the TIS framework could be used to address the third question by identifying opportunities and challenges (referred to in the TIS literature as inducement and blocking mechanisms), which in turn could be used to form recommendations for policymakers and managers. 17 1.2 Aim This master’s thesis aims to establish new knowledge and a better understanding of the development of hydrogen valleys – a multi-stakeholder innovation system centered around a disruptive new technology. This aim will be addressed by mapping the technological innovation system of a newly founded hydrogen valley called Mid Sweden Hydrogen Valley. 1.3 Delimitations This master’s thesis is made on a system level, and thus, detailed descriptions of technical and chemical components related to hydrogen valleys are left out. Due to restrictions on resources and access, not all actors in Mid Sweden Hydrogen Valley are interviewed. Since the thesis is a case study of MSHV, there are also limitations in terms of generalizability for hydrogen valleys on a more global level. Finally, since this study employs a research method that sometimes involves subjective measurements, such as when evaluating the performance of the hydrogen valley, there may be inherent biases that could potentially limit the objectivity of some findings. 1.4 Specification of the issue being investigated Based on the background, aim, and limitations, three research questions are presented: RQ1: What are the structural components, i.e. actors, networks, institutions and infrastructure, in a hydrogen valley and how are they interconnected? RQ2: How do the functional dynamics affect the development and performance of a hydrogen valley? RQ3: Which inducement and blocking mechanisms are present in a hydrogen valley, and how do they impact the valley’s development? 18 2. Theory In this section, relevant literature on technological innovation systems is presented, followed by descriptions of the structural components, functions, and inducement and blocking mechanisms. In addition, the last section presents critique of the TIS literature. 2.1 Introduction to technological innovation systems The analysis of innovation processes and socio-technical transformations is one of the classical research fields within the innovation literature. The creation of innovation processes to accelerate innovation is crucial, as it significantly impacts economic growth, societal welfare and the environment (Hekkert et al., 2007). While the current use of technologies often has a negative impact on the environment, development of sustainable technologies is important to reduce the environmental footprint. Sustainable technologies alone do not drive sustainable technological change; rather, they need to be aligned with the social perspective, including user behaviors, regulatory frameworks and industrial networks (Hekkert et al., 2007), which explains why studying socio-technical change is a complex process. Markard & Truffer (2008) argue that, even though socio-technical transformations span many decades of research, they lack understanding of the complex underlying innovation processes – which often include a myriad of stakeholders throughout the value chain. The TIS literature was developed to advance the comprehension of these innovation processes and societal transformations, by providing holistic frameworks that explore the dynamics and interactions between stakeholders in an innovation system (e.g., see Bergek et al., 2008a; Markard & Truffer, 2008; Hekkert et al., 2007). TIS has reached widespread diffusion among innovation scholars (Bergek, 2019), where one of the most prominent papers is Bergek et al. (2008a). The paper presents a framework for mapping the structural components of the TIS and assessing the performance by analyzing the dynamics of the key functions that influence the development, diffusion and use of the new technology (i.e. in the form of an artifact or knowledge). As mentioned in the introduction, a TIS is defined as: “a network of agents interacting in the economic/industrial area under a particular institutional infrastructure and involved in the generation, diffusion and utilization of technology” – (Carlsson and Stankiewicz, 1991) The ultimate goal of TIS analyses is to outline the challenges and opportunities that exist within the TIS and in turn provide policy recommendations that mitigate and promote these respectively (Bergek et al., 2008a). If new and innovative TISs are to be developed successfully, Kivimaa & Kern (2016) argue 19 that policies need to be in place that enable creative destruction of the old and competing TISs. For energy transitions, this implies having policies that for example disincentivize the use of fossil fuels. Their key argument is that sustainability transitions should have a mix of two kinds of policies: the first ones being the policies that incentivize the niche-innovations (i.e. less established innovations) and building effective innovation systems around them, and the second ones being the policies aimed at destabilizing the current dominant regimes (i.e. the more established technologies and practices) and creating openings for the take-off and sustained growth of niche innovations to replace the incumbent technologies. 2.2 Constituents of technological innovation systems TIS analyses are differentiated between structural components and functional dynamics, which are two complementary ways to describe how a TIS works. These two constituents are described below. Additionally, inducement and blocking mechanisms explain why the TIS works the way it does, which will also be presented. 2.2.1 Structural components The structural components refer to actors, networks and institutions (Bergek et al., 2008a; Hekkert et al., 2007). A fourth structural component is proposed by Wieczorek & Hekkert (2012), that also adds infrastructure. Mapping the structural components creates a more in-depth understanding of how the TIS works, and also provides the foundation for the functional analysis. Below, each of the structural components are introduced. Actors Actors are the organizations that in some way contribute to the emerging technology, either directly as a developer or adopter, or indirectly as a financer or regulator (Suurs, 2009). Since it is the actors who are creating, diffusing and using the technologies, the development of a TIS is dependent on actors' presence, skills, activities, and collaborations with other actors. The latter could be explained by, for instance, grant providers not knowing where financial support is needed if developers do not provide them with information (Suurs, 2009). Actors include companies, government and non-governmental agencies, research facilities, universities, venture capitalists, and associations (Bergek et al, 2008; Markard & Truffer, 2008). Networks Networks could be defined as a particular group of actors with strong linkages (Suurs, 2009). Networks describe how and where communication and interactions take place and can either be formal or informal 20 (Bergek et al, 2008). Musiolik et al. (2012) describes formal networks as an organizational structure with easily identified members that collaborate to achieve common aims or to solve specific tasks. The authors argue that formal networks are important in the development of TIS as they facilitate coordination of strategies and enable collective action among participants (Musiolik et al., 2012). According to Bergek et al. (2008), informal networks are more difficult to identify and often require discussions with industry experts or other actors. This process is iterative, with additional information being incorporated as the analysis progresses. Furthermore, networks play a crucial role in the development of TIS because they facilitate exchange of knowledge, fosters a learning process, and create synergies between actors (Suurs, 2009). Institutions Institutions are described as the culture, norms, laws, regulations, and routines that need to be aligned for the technology to diffuse (Bergek et al, 2008). Suurs (2009) makes a distinction between formal and informal institutions, where the formal are rules set by some authority, and informal are tacit and developed organically through collective interactions of actors. The institutional rules for a TIS in a formative phase are often underdeveloped and typically not suitable for the emerging technology. According to Suurs (2009), visions and expectations are the primary reasons for supporting an emerging technology. Infrastructure Infrastructure as a structural component is often neglected in the TIS literature, and there is no common consensus about what infrastructure covers. Wieczorek & Hekkert (2012) takes a broad perspective on infrastructure and divides it into three categories, including physical, financial and knowledge. The physical infrastructure for example covers buildings, machines, and power grids. The financial infrastructure covers subsidies and grants, while knowledge infrastructure includes expertise and know- how. Furthermore, Smith (1997) argues that physical infrastructure significantly influences the establishment of technological dominance and the forming of technological trajectories, which impacts the performance of an innovation system. 2.2.2 Functional analysis Having addressed the structural components, the second major part of the TIS framework is about analyzing seven key functions, to determine the extent to which the functions are currently filled in that TIS (Bergek et al, 2008). A well-performing system should ideally have all functions fulfilled at a satisfactory level. However, this could provide a challenge for analysts, since there is no exact way to appraise satisfactory performance. How the performance of each function is appraised is presented later in the method chapter. 21 The key functions are: (1) knowledge development and diffusion, (2) influence on the direction of search, (3) entrepreneurial experimentation, (4) market formation, (5) legitimation, (6) resource mobilization, and (7) development of positive externalities. The identification of functions complements the structural analysis, since it emphasizes what the system does and how it works in comparison to how it is composed or structured (Bergek et al., 2005). However, structure and function are two intertwined sides of the same object; the system, since functions influence the system structure and vice versa (Markard & Truffer, 2008). Even though structure and functions are intertwined in a system, functions are still an important indicator of system performance (Markard & Truffer, 2008), and therefore important to analyze separately. Knowledge development and diffusion Knowledge development is essential for the development of new technologies, and involves learning activities, which range from basic science to learning by practicing (Suurs, 2009). Hekkert et al. (2007) distinguishes these activities as “learning-by-searching” and “learning-by-doing”. Suurs (2009) argues that universities and research institutes have the main responsibility for knowledge development, but that other actors often contribute with “learning-by-doing”. Furthermore, there are different types of knowledge, such as scientific, technological, production, market, and design – which all could take various forms; e.g. R&D, learning from new applications, production, and imitation (Bergek et al., 2008a). While knowledge development is a central function in a TIS, its tendency to create variety causes uncertainty in the system, which is mitigated by the two other functions of influence on the direction of search and entrepreneurial experimentation (Suurs, 2009). Knowledge diffusion is closely related to networks, considering the exchange of knowledge between the actors in a network (Suurs, 2009). Usually, knowledge diffusion occurs in partnerships between actors and in meetings like seminars and conferences. While there is a tendency that actors within a certain community shares more knowledge than actors from different communities, it is nevertheless crucial that there is knowledge diffusion between actor groups for the development of a TIS (Suurs, 2009). Technology developers, for example, need to provide policymakers with information to allow them to establish suitable policies. It is therefore important that knowledge is shared widely between actors in a TIS. The knowledge development and diffusion function is commonly mapped through R&D projects (quantity and size of investments), number of patents, learning curves (Bergek et al., 2008a; Hekkert et al., 2007), bibliometrics, professors, and assessments by managers (Bergek et al., 2008a). 22 Influence on the direction of search Influence on the direction of search, also called guidance of the search by Hekkert et al. (2008) and Suurs (2009), refers to the incentives actors have to enter a TIS and also to the factors that influence the direction of search within a TIS (Bergek, 2008). Hekkert et al. (2008) relates the latter to the fact that resources are always limited and that a selection has to be done to allow a sufficient amount of resources to each option. This also explains how this function reduces the uncertainty created in the knowledge development and diffusion function which was described above. However, Suurs (2009) points out that there is a fine balance between creating and reducing variety, because too much focus could lead to a lack of variety. The author further concludes that the selection in this function is important to create a clear sense of direction regarding the technology. Otherwise, efforts within the functions knowledge development and diffusion and entrepreneurial experimentation are unlikely to yield meaningful results. Since influence on the direction of search also involves incentives that actors have to join a TIS, the function is also important in attracting new entries to the TIS (Suurs, 2009). Influence on the direction of search function is commonly influenced through an interactive process where actors share ideas and knowledge about the technology (Hekkert et al., 2007). This forms visions and expectations that steer the direction of the innovation system. In case an actor shows success within some specific area, this often has a large impact on the direction of search (Hekkert et al., 2007). The influence on the direction of search function is often indicated by expectations on growth potential (Bergek et al., 2008a; Hekkert et al., 2007), economic incentives (e.g. taxes, subsidies etc.), regulatory pressures, interest from leading customers (Bergek et al., 2008a) and targets set by governments and industries (Hekkert et al., 2007). Entrepreneurial experimentation Entrepreneurial experimentation involves testing of new technologies and applications (Bergek et al., 2008a). Entrepreneurs have a central role in the development of a TIS and are responsible for transforming the potential of new knowledge, networks and markets into practical actions that create new business opportunities (Hekkert et al., 2007). Since experimenting is a way to learn more about how technologies are functioning under different circumstances, it creates a learning process that reduces the uncertainty within a TIS (Hekkert et al., 2007). The development of emerging technologies is often unpredictable because they are not aligned with the existing structural components. However, through continuous experimentation and adaptation, these technologies could gradually be adjusted to fit its structural environment, and the structural components can be shaped to better support the emerging technologies (Suurs, 2009). 23 An entrepreneur can be either a new entrant contributing to the formation of a new market, or an incumbent company diversifying its current business to influence and capitalize on new developments (Hekkert et al., 2007; Suurs, 2009). When analyzing this function, it is therefore important to map the number of new entrants and diversifying incumbent companies (Bergek et al., 2008a; Hekkert et al., 2007), the number of different types of applications, the breadth of technologies used (Bergek et al., 2008a), the number of diversification activities of incumbent actors, and the number of experiments with the new technology (Hekkert et al., 2007). Market formation For an emerging TIS, a market may be very limited or not exist at all (Bergek et al., 2008a). The authors describe that market formation goes through three phases, including “nursing market, bridging market, and mature market”, and it is important to understand in which phase the specific TIS is in to give a fair assessment of the function. A nursing market indicates that a TIS is in a formative phase, while a bridging market indicates that a TIS is in a growth phase (Bergek et al., 2008b). Furthermore, it is often difficult for new technologies to compete with existing ones, due to their weaknesses (Hekkert et al., 2007), such as for example price/performance gap to incumbent technologies (Jacobsson & Johnson, 2000). This will lead to that diffusion will be slow and the new technologies will require some sort of support. The market formation function therefore involves activities that contribute to the creation of demand for an emerging technology, including for instance financial support or higher taxes on existing technologies (Hekkert et al., 2007; Suurs, 2009). Hekkert et al. (2007) also argue that forming temporary niche markets for specific applications is a possible solution to the challenges that new technologies are facing, since actors could learn more about the technology and form expectations on the technology. To analyze the market formation, it is important to identify the market development and what drives the market formation (Bergek et al., 2008a). To accomplish this, the analyst could map what phase the market is in (i.e. nursing, bridging, or mature), market size, customer behaviors, institutional stimuli for market formation (Bergek et al., 2008a), number of niche markets, tax regimes and environmental standards for new technologies (Hekkert et al., 2007). Legitimation Legitimation describes the social and industrial acceptance for the underlying technology. In order for a TIS to perform optimally, the underlying technology needs to be considered appropriate and desirable by a majority of actors (Bergek et al., 2008a; Hekkert et al., 2007). This legitimation is essential for demand to form, for mobilization of resources to occur, and for political strength to form in the TIS 24 (Bergek et al, 2008). In addition, legitimation also affects managers expectations and hence the influence on the direction of search function. This function can be analyzed by mapping interest groups and their efforts to influence policy and decision-makers (Hekkert et al., 2007), stakeholders’ social acceptance and activities within the TIS that could increase the legitimacy (Bergek et al., 2008a). This will increase the understanding of the relative strength of the legitimacy, what influences the legitimacy and how it influences the market and legislation (Bergek et al., 2008a). Resource mobilization Resource mobilization emphasizes the necessity for different kinds of resources to be allocated for the successful development of a TIS. The different kinds of resources are: (1) human capital (e.g. technical competence and managerial competence), (2) financial capital, and (3) complementary assets (e.g. complementary products, services, and network infrastructure) (Bergek et al., 2008a). This function can according to Bergek et al. (2008a) be analyzed by “identifying rising volume of capital, increasing volume of seed and venture capital, changing volume and quality of human resources, changes in complementary assets”. However, Hekkert et al. (2007) argue that it is difficult to map this function by specific indicators and instead propose that the best way is to through interviews ask actors if they perceive access to the resources as problematic or not. Development of positive externalities Development of positive externalities acts as a reinforcement mechanism of the other six functions rather than as an independent function (Bergek et al., 2008a). According to Suurs (2009), the positive feedback loops, resulting from interaction between functions, are crucial in the development of a TIS. For example, as more and more actors enter the TIS, legitimation may strengthen, which in turn creates a virtuous circle where the functions resource mobilization, influence on the direction of search, market formation, and entrepreneurial experimentation, continuously strengthens. In addition to more actors entering the TIS, development of positive externalities could be strengthened by mechanisms such as resolution of uncertainties, combinatorial opportunities, pooled labor markets, specialized intermediates, and knowledge spillovers (Bergek et al., 2008a). 2.2.3 Inducement and blocking mechanisms A TIS in an early phase usually shows weak functional dynamics and develops slowly, since the environment tend to favor established TISs (Bergek et al., 2008a). The reasons for this partially derive from the underdeveloped structural components, but also from the larger context that surrounds the TIS (Bergek et al., 2008a). Hence, the third major part of the TIS framework is to identify mechanisms that 25 either promote or inhibit the development of the functions in a TIS (Bergek et al., 2008a; Johnson & Jacobsson, 2001). These mechanisms are referred to as inducement and blocking mechanisms. Previous studies by Johnson and Jacobsson (2001) show that most inducement mechanisms within the renewable energy technologies field derive from governmental policies. For example, these include funding programs, investment subsidies, and policies aimed at stimulating market formation by changing relative prices. In addition, the authors also highlight general environmental concern as a factor that stimulates the market formation of renewable technologies. Jacobsson and Johnson (2000) argue that the inducement mechanisms need to be strong enough to overcome the weaknesses of a new technology. Moreover, Johnson and Jacobsson (2001) discuss several blocking mechanisms within the same field. The first blocking mechanism is related to the weak price/performance ratio that new technologies have compared to incumbent technologies. Incumbent technologies typically benefit from economies of scale making them more cost-effective. Secondly, the weak and underdeveloped relationships between actors within a new TIS create another blocking mechanism. This mechanism results in slow knowledge development and diffusion and affects the influence on the direction of search negatively (Johnson & Jacobsson, 2001). A lack of communication between actor groups also implies uncertainty regarding the demand, which in turn inhibits the market formation. Furthermore, governments commonly lack a clear vision with long-term goals regarding sustainable technologies, creating a third blocking mechanism. Other identified blocking mechanisms include lock-in to incumbent technologies, customers lack knowledge to invest in the technology and articulate demand and firms’ limited influence on policy-making (Johnson & Jacobsson, 2001). 2.3 Critique Authors such as Coenen et al. (2012) address the issue of focusing too much on the functional domain, which they critique Bergek (2008a) for doing. Their argument is that every TIS is unique spatially and that too much focus on the functional domain leads to overgeneralization. In other words, just because a system performs well in one economic geography does not mean that the same functions will perform as well in other economic geographies. The economic geography does not only emphasize the physical geography, i.e. that actors are close in proximity to each other, but also include the non-physical connections, e.g. how the actors cooperate and communicate with each other. This argument emphasizes the importance of thoroughly mapping and analyzing the structural components, since it creates a better understanding of the TIS and results in more meaningful findings. Bergek et al. (2015) responded to the critique from Coenen et al. (2012) by complementing their previous framework with four additional 26 structural considerations: (1) interaction between a focal TIS and other TISs, (2) interaction between a focal TIS and relevant sectors, (3) TIS development in geographical context structures, and (4) interaction between a focal TIS and the political context. Another criticism that has been put forward by Wieczorek & Hekkert (2012) is that infrastructure is a critical aspect of many TISs, but it is not included in the Bergek et al. (2008a) framework. These criticisms are incorporated into the method that is used to analyze MSHV, which is described in the next chapter. 27 3. Method This thesis employs a qualitative research strategy. As described by Bell et al. (2019), qualitative research is concerned with words rather than numbers and a constructionist ontological position, meaning that reality is based on the subjective interpretation made by individuals rather than there being a truth “out-there” in a fixed state. This is a favorable strategy when applied to social sciences (as opposed to natural sciences), since answers are often ambiguous and situational (Franzosi, 1997), and therefore hard to quantify. The data gathered in the study of hydrogen valleys are of an ambiguous and non-quantifiable nature, e.g. due to the novelty and explorative character of a hydrogen valley, which therefore requires a qualitative research strategy. Furthermore, our research applies an abductive approach, which can have the advantages of both deductive and inductive approaches while simultaneously overcoming the weaknesses associated with the respective approach (Bell et al., 2019). Our research has deductive components, since preconceived theories, such as the framework outlined by Bergek et al. (2008a), are used in a relatively sequential and linear way. However, we have remained open about revising theory, scope, and research questions as the research process unfolded, which in contrast indicate an inductive approach. Therefore, the research approach could be described as abductive, which is a favorable approach in this case because of two main reasons: firstly, the existing framework developed by Bergek et al. (2008a) provides a good basis for analyzing the dynamics of a hydrogen valley, and secondly, new insights from the data collection could simultaneously be used to make alterations to the framework and to realign the research objectives to yield more fruitful findings. The research design is a case study, meaning that one specific organization, institution, location, person, or event is analyzed in depth, which therefore provides detailed and exact findings about that particular case (Flick, 2014). The case that is investigated is MSHV. One of the main disadvantages of the research design is that generalizability could be difficult when only focusing on a single case (Bell et al., 2019), which could make further theory-building problematic. However, rather than pursuing a sample-to-population logic, case studies have been shown to provide analytical generalizability, where findings can apply between cases of different natures (Yin, 2013). In addition, we have addressed the disadvantage of case studies in two ways: firstly, by approaching the research with caution, acknowledging the limitations inherent in the case study approach, and secondly, by avoiding overgeneralization – rather than asserting universal truths, articulating that findings are likely applicable to other cases. 28 An overview of the research methodology can be seen in Figure 2 below. The arrow pointing backwards highlights that minor aspects of the study have been revised and iterated as the process unfolded. In the sections that follow, each step of the methodological process as outlined in Figure 2 is presented. Figure 2. Overall research methodology. The dotted lines showcase the abductive approach. 3.1 Framing research The first step of the research process was to frame the research, which was done by formulating research objectives and reviewing the existing literature. The research objectives were primarily based on an exploratory background search, which included both internal documents and internet searches. The preliminary research objectives were then formulated – including a preliminary description of aim, limitations, and research questions, and later revised several times as the writing process unfolded. After the research had been framed, the next step was to review the literature. The literature review was conducted for three main reasons: (1) to get acquainted with the subject and learn about relevant theories, (2) to narrow the research scope in order to fit with existing theories, and (3) to provide a guideline for the analysis. The different theoretical fields that were investigated are presented in Table 1 below. As showcased, the literature search started off much wider, encompassing different theoretical fields, e.g. innovation hubs, technological transformations, energy business models, but ultimately narrowed down to the TIS literature as the primary literature source. 29 Search terms Titles read Abstracts read Articles skimmed Articles read thoroughly Hydrogen valley 30 9 9 6 Innovation hub, industrial clusters 220 20 20 3 Hydrogen value chains 65 7 3 2 Technological transformations 145 12 9 1 Energy business models 200 34 19 9 Technological innovation systems 45 24 20 14 Table 1. Articles in the literature search. 3.2 Data collection The second step in the process was the data collection. Interviews were the main method for gathering data, which was complemented with secondary data collection. Since this thesis applies a qualitative research strategy, the data collection did not include many of the quantitative measurements suggested by Bergek et al. (2008a) and Hekkert et al. (2007). Another reason for not including quantitative measurements, such as number of patents when analyzing knowledge development and diffusion, is that we study a regional TIS in a formative phase where this type of data is limited or restricted, making it difficult to analyze. 3.2.1 Interviews The primary method for gathering data has been through interviews with different actors and stakeholders in MSHV. The main reason for using interviews as the primary data collection method was that nearly all actors within the TIS could be interviewed, which could lead to comprehensive and relatively exhaustive results. Interviews in qualitative research are generally open to allow for flexibility (Bell et al., 2019), and since our research questions are of a qualitative nature, this openness has been employed in the interviews. More specifically, the interviews have been semi-structured, where topics or questions were prepared in advance, but where the interviewee had great flexibility in how they replied. Since our research is of exploratory character, semi-structured interviewing allowed the actors 30 to give more nuanced answers that would otherwise not have been possible in a stricter interview format. In addition, this allowed us to ask follow-up questions based on the interviewees’ replies if we needed clarification or found a new and interesting perspective to pursue. In turn, this allowed for a deeper analysis and better preparedness to answer the research questions. To gain answers from different perspectives, interviewees were chosen from different actors along the value chain in MSHV. In order to increase the reliability of the answers, at least 2 interviews were conducted within each actor group in the value chain (including renewable energy providers, hydrogen producers, storage/distribution providers, hydrogen users, grant providers, incubators and research institutes, authorities, and OEMs). See Table 2 for an overview of all the interviews (without a mapping of the actor groups – this is done in the analysis chapter). The sampling method used was purposive sampling, as described by (Bell et al., 2019), since the interviewees were chosen strategically to improve variety and quality of the answers. The interview questions were centered around the research questions and the TIS framework, i.e. by asking about the actors, networks, institutions, functional dynamics, and the challenges and opportunities in the hydrogen valley. An online video format was used in the interviews, since it is more flexible and less time- and cost-intensive compared to face-to- face interviews, especially given the geographical distances involved. Video interviewing also creates a similar interactional experience as face-to-face interviewing (Bell et al., 2019). The interview participants were predominantly senior executives or other employees with particular insight of the organization’s sustainability strategy. 31 Table 2. Interviews The interview structure was set by an interview guide, which Bell et al. (2019) describes as a set of memory prompts that the interviewer can use to remember all areas that should be covered during the interview. We did not follow the interview guide strictly – instead, we allowed for flexibility in the interview process, e.g. by asking follow-up questions and investigating inconsistencies in the answers to find important nuances. Since interviews are the main data collection method for this thesis, a thorough approach to interviewing was adopted. The process of transcription was used, since it can counteract the limitations of memory and biases, as well as allow for repeated examinations of the interviewees' answers (Bell et al., 2019). To aid in this process, interview sessions were recorded and inputted into transcription tools. Naturally, a prerequisite for this step is that the interviewees gave consent to being recorded. In addition to transcription, note-taking of key points was adopted during the interviews. As a last step in the interview process, segments in the analysis, which included specific information about the interviewee or the organization, were sent out after it was written for two distinct purposes: (1) to allow the interviewee to comment on the accuracy of the segment and give suggestions, and (2) to ask the Organizations interviewed Alleima Nitiu Region Gävleborg (MSHV coordinator) Nordion Energi University of Luleå (Researcher) Ovako Dalavind Plagazi Energimyndigheten Powercell Green Iron Skyborn Renewables Gävle Harbor Volvo Group Hydri Sandbacka Science Park Klimatklivet Sandvik Länsstyrelsen Statkraft MaserFrakt Total interviews: 21 Average length: 48 min 32 interviewee of permission to include it in the report by not accidentally publishing sensitive or confidential information. 3.2.2 Secondary data collection According to Bell et al. (2019), documents are often used as a supplementary data collection method to semi-structured interviews in a case study, since it can highlight important past managerial decisions and actions. The type of documents that was used are for example reports from different agencies, companies, or stakeholders in MSHV. Different news articles were also used from reputable sources. According to Bell et al (2019), it is important to consider that documents are written with a distinctive purpose in mind which does not always reflect reality. Therefore, to assess the quality of documents used in this research, four criteria have to be fulfilled, as suggested by Bell et al (2019): (1) authenticity, (2) credibility, (3) representativeness, and (4) meaning. Hence, the secondary data sources that were used were qualitatively evaluated based on these four criteria. The main purpose of the secondary data collection was to verify important factual information from the interviews and also to explore certain topics more in depth that were brought up in the interviews. 3.3 Analysis An important question to answer is when to stop collecting more data and instead focus on theory building and drawing conclusions from the data. According to Bell et al. (2019), there comes a point in the data collection process where new data are no longer illuminating the themes further. This concept is called theoretical saturation, which was used to determine the point in time when further data collection was no longer beneficial. When theoretical saturation was reached, the first step was to do a thematic analysis of the data, by grouping the data into themes. Secondly, theory found in the literature search, e.g. from Bergek et al. (2008a), was used to analyze the themes. The analysis process, including the thematic analysis and the analysis based on the TIS theory, are described below. 33 3.3.1 Thematic analysis For the analysis of the data, a thematic analysis was used. Bell et al. (2019) describes thematic analysis as one of the most common forms of qualitative data analysis, which is used to group data into themes before analyzing it further. When looking for themes, it is important to look for repetitions of topics, specific typologies, or similarities and differences between topics (Bell et al., 2019). Themes were found throughout the whole data collection process, and the themes centered around our research questions and the TIS framework, see Figure 3. 3.3.2 Steps in the analysis of technological innovation systems Bergek et al. (2008a) presents a framework including six sequential steps that should be taken when analyzing a TIS. These include: (1) defining the TIS, (2) identifying structural components, (3) mapping the functional patterns of the TIS, (4) assessing the functionality of the TIS and setting process goals, (5) identifying inducement and blocking mechanisms, and (6) specifying key policy issues. The steps will be described in more detail below. The first step is about deciding the scope of the TIS. Bergek et al. (2008a) mention three choices in this step: (1) the choice between knowledge field or product as a focusing device, (2) the choice between breadth and depth, and (3) the choice of the spatial domain. Depending on the research questions – focusing on the product (e.g. a hydrogen fueling station) or on an entire knowledge field (e.g. hydrogen as a technology) is the first choice. The second choice refers to the degree of specificity, i.e. if the analysis focuses on one application and studies it in depth – or if a wide range of applications are studied. The last choice of the spatial domain refers to whether the study is limited to a geographical area such as a hydrogen valley or if it takes a global perspective. These three choices have implicitly been presented in the background, which could be presented explicitly as (1) low-carbon hydrogen as a knowledge field – as the case study covers many use cases of hydrogen and not just a specific product, (2) a broad focus made on a system level, and (3) the spatial domain being MSHV – a regional hydrogen valley in mid-Sweden. Figure 3. Themes in the thematic analysis 34 Figure 4. The six steps in the analysis of a TIS. From Bergek et al. (2008a). The second step is about mapping the actors, networks, and institutions of the TIS (Bergek et al., 2008a). The mapping could for example be done through industry associations (e.g. through company documents) and interviews with experts. If the TIS is in an early formation phase like MSHV, structural components may be weak and hard to identify (Bergek et al., 2008a). According to the authors, this requires an iterative process where additional pieces of information are added as the analysis proceeds, e.g. by adding some informal networks that are often hard to identify initially, later in the analysis process. This step, together with all remaining steps, are based on the data collection, and presented in the analysis chapter. The third step is used to understand how the TIS works rather than how it is structured, by analyzing a set of processes, which Bergek et al. (2008a) label as functions, that are important in the development of the TIS. The functions have been described in the theory chapter earlier. The purpose of this step is twofold; one part being the outlining of the functions and the second part being the appraisal of each function's strengths and weaknesses (as highlighted as 3a and 3b in Figure 4). The outlining is presented through a general description of what affects the function in the TIS, while the appraisal of strengths and weaknesses places more weight on how the findings affect the development of the TIS, i.e. in a positive or negative way. The fourth step in the analysis is about assessing the functionality of the TIS and setting process goals, i.e. assessing the performance of the TIS based on the result of the functional analysis. However, this can prove problematic, since it is difficult to objectively assess the “goodness” of a particular function (Bergek et al., 2008a). Although the analysis of the functions in the previous step is important, it is not 35 sufficient to determine if the TIS is well performing or not, since a function that is weak does not necessarily constitute a problem, and a function that is strong is not always an important asset. Bergek et al. (2008a) describe two ways of assessing the performance: (1) by identifying the phase of development of the TIS and (2) through comparisons between TISs. Depending on which phase of development the TIS is in (i.e. formative or growth phase), functions will need to be evaluated differently. For example, a common mistake is that a TIS within the formative phase will be evaluated based on criterias which are more suitable for a TIS in a growth phase (Bergek et al., 2008a). If a TIS in a formative phase is analyzed without considering that this phase is characterized by small volumes and slow diffusion, the TIS could inappropriately be identified as a failure even though the TIS may be very effective in reality. The analyst has to assess the functions based on the appropriateness of that function's performance in that particular phase. The formative phase is characterized by Bergek at al. (2008a) by the following: ● the time dimension, where formative periods often extend beyond a decade; ● large uncertainties in technologies, markets and applications; ● price/performance of the products being underdeveloped; ● a volume of diffusion and economic activities that is but a fraction of the estimated potential; ● absence of powerful self-reinforcing features (positive feedbacks) and weak positive externalities. The second way to assess the functionality of the TIS and set performance goals could be made by comparing different TISs. These TISs should be similar or related to the focal TIS, and their performance should be evaluated to gauge what performance is reasonable for the focal TIS (Bergek et al., 2008a). Bergek et al. (2008a) suggest using phase analysis and comparative analysis in combination to reach a conclusion about the performance of the TIS and to set new functional goals. New functional goals may be to broaden the knowledge base or widen the range of experiments (if the functions knowledge development and diffusion and entrepreneurial experimentation are identified as weaknesses). It is important that the functional goals are not expressed as final goals – e.g. that a TIS in a formative phase should achieve higher growth – since this is not an indicator of functionally poor performance. While these two ways showcase how the performance of the TIS could be evaluated, it is important to note that this evaluation is inherently qualitative and subjective. There are currently no “objective” ways to measure performance to our knowledge, nor are there any indications of this in the TIS literature. This is a significant limitation of the TIS framework. However, by using both these ways, together with the isolated analysis of the function, a relatively comprehensive overview could be given, which should give a very adequate appraisal of the function’s performance. 36 The evaluation of the functions is assessed on a three-point ordinal scale, including weak, intermediate and strong (as done by previous TIS analyses). When there are clear arguments indicating that a function is performing poorly with respect to the TISs stage of development, the function is graded weak. Similarly, a function that has clear arguments showing good performance is graded strong. In case the arguments are ambiguous, and the function cannot be graded either weak or strong, the function is graded intermediate. Moreover, it is important to note that the actual grade a function is given is based on a holistic perspective, which implies that a function that is for example graded weak, could still have some strong components and vice versa. Step five relates to finding key inducement mechanisms (opportunities) and blocking mechanisms (challenges) for the TIS. These can be both internal (i.e. related to the functions of the TIS) or external (i.e. not directly related to the TIS). An example of an external mechanism is global warming, which could constitute either as an inducement mechanism or as a blocking mechanism, depending on the TIS. The internal mechanisms are related to the functional analysis of the TIS (step 3). Lastly, the sixth step of the TIS analysis is about specifying key policy issues based on the previous five steps. Policy should ideally be aimed at improving TIS performance by (1) strengthening/adding inducement mechanisms and (2) weakening/removing blocking mechanisms (Bergek et al., 2008a). In Figure 5 below, an example of the result from the analysis in step 1-6 is presented for the case “IT in home care” (made by Bergek et al., 2008a). 37 Figure 5. Example of findings from the TIS analysis. Results from step 1-6 within the case “IT in home care”, made by Bergek et al. (2008a). The function of positive externalities is indicated by the dotted lines between the other functions. 3.3.3 Our analytical framework The steps presented in the previous section is the core analysis, which is applied to the case of Mid Sweden Hydrogen Valley. However, some deviations from the steps presented in Bergek et al.’s (2008a) framework are made with the purpose of better answering the research questions. One of these deviations is in the second step when mapping the structure: we complement the Bergek et al. (2008a) framework by adding infrastructure as a fourth component, as suggested by Wieczorek & Hekkert (2012). However, we only choose to include the physical infrastructure, since financial and knowledge infrastructure are included in the other functions by Bergek et al. (2008a). We also complement the framework by including four additional structural considerations: (1) interaction between a focal TIS and other TISs, (2) interaction between a focal TIS and relevant sectors, (3) TIS development in geographical context structures, and (4) interaction between a focal TIS and the political context, in an attempt to make the findings more generalizable to other TISs, as suggested by Bergek et al. (2015). Additionally, since the blocking and inducement mechanisms for the TIS are both internal and external in nature, these findings are just as fruitful for managers operating within the TIS as it is for policymakers – thus justifying managerial implications of these findings as well. The analytical framework, including notable deviations from the steps proposed by Bergek et al. (2008a), are presented in the Table 3 below. 38 Research questions Theoretical frameworks (main framework – Bergek et al., 2008a) RQ1: What are the structural components, i.e. actors, networks, institutions and infrastructure, in a hydrogen valley and how are they interconnected? Step 1: Defining the TIS Step 2: Structural analysis Adding infrastructure as a fourth structural component (as suggested by Wieczorek & Hekkert, 2012) Adding four additional structural considerations: (1) interaction between a focal TIS and other TISs, (2) interaction between a focal TIS and relevant sectors, (3) TIS development in geographical context structures, and (4) interaction between a focal TIS and the political context (as suggested by Bergek et al., 2015) RQ2: How do the functional dynamics affect the development and performance of a hydrogen valley? Step 3: Functional analysis Step 4: Assessing the performance of the TIS RQ3: Which inducement and blocking mechanisms are present in a hydrogen valley, and how do they impact the valley’s development? Step 5: Blocking mechanisms – both from internal analysis (expand from structural and functional) and external analysis. Step 6: Implications for policymakers Adding managerial implications. Table 3. Analytical framework 3.4 Reflection on validity The term validity refers to the trustworthiness of the conclusions drawn in the study, or in other words, if the study accurately measures what it claims to measure (Bell et al., 2019). Using an empirically validated and systematic approach, such as the TIS framework, should improve the validity of the study. However, some of the inherent limitations of the TIS framework is that the analytical steps are evaluated subjectively, meaning that the analysis can be influenced by biases and misconceptions, which risks decreasing the validity of the study. However, we have attempted to minimize this limitation in three main ways: (1) by relying on a robust foundation of interview data and reliable secondary sources for analysis (2) by being critical towards the interview data – realizing that the interviewees can themselves have biases and misconceptions, and (3) by double checking important or speculative information. It is also important to mention that the study was not conducted for MSHV, but rather for an independent company (CIT Renergy), thus minimizing potential conflicts of interest. 39 4. Analysis The analysis is presented in four sections: (1) structural mapping, (2) functional analysis and performance appraisal, (3) inducement and blocking mechanisms, and (4) implications for policymakers and managers. 4.1 Structural mapping The structural mapping includes a description of the actors, networks, institutions, and infrastructure that are either within MSHV or that have a close connection to it. The purpose of the structural mapping is twofold. Firstly, it can lead to interesting insights regarding the connections within MSHV. Secondly, it serves as a basis for the functional analysis by providing the characteristics of the hydrogen valley, which in some cases could explain why the functions perform the way they do. In order to present the structural mapping in a coherent and clear way, the actors have been categorized into different actor groups. The overall connections between the actor groups are illustrated in the hydrogen value chain seen below in Figure 6. The figure includes the main hydrogen value chain, and also facilitators and producers of equipment as two supporting parts. Based on the interviews and previous studies that define hydrogen value chains (e.g. Masip et al., 2021), the main hydrogen value chain moves from energy provision, to hydrogen production, to storage and distribution, and finally to the end users. There is a logical flow of energy throughout the whole value chain, but it is important to notice the energy shift from electricity to hydrogen between energy provision and hydrogen production. The facilitators include grant providers, incubators and research institutes, and authorities. These play an important role in supporting the hydrogen value chains through financial capital, R&D and legal support, especially in this early stage where the market is still underdeveloped. The producers of equipment are another crucial set of actors that are enabling the creation of a hydrogen value chain, by for example manufacturing production equipment and fuel cells that are used in the main hydrogen value chain. However, it is important to note that neither the facilitators nor the producers of equipment are directly linked to the energy/hydrogen flow between the actor groups in the value chain, and therefore are illustrated outside of the main value chain. 40 Figure 6. Hydrogen value chain 4.1.1 Actors In this section, a mapping of the actors is presented, including each actor’s connection to MSHV. The actors consist of members of MSHV and other influential entities contributing to the development of MSHV. They are classified into actor groups corresponding to the different stages of the hydrogen value chain (Table 4). Note that certain actors might fit into more than one actor group, depending on their involvement across various parts of the value chain. 41 Actor groups Actors interviewed Actors not interviewed Renewable energy providers Skyborn Renewables, Dalavind, Statkraft Svea Vind Offshore, Vattenfall Hydrogen producers Ovako, Plagazi, Statkraft Linde Gas Storage/Distribution Nordion Energi, Hydri, Nitiu Inlandsbanan, Hynion Hydrogen users Ovako. Alleima, Sandvik, MaserFrakt, Green Iron Outokumpu, Inlandsbanan Grant providers Energimyndigheten, Naturvårdsverket Tillväxtverket Incubators and Research institutes Sandbacka Science Park, Hydrogen researcher at Luleå University, Gävle Harbor Dalarna Science Park, RISE, Mellansvenska Handelskammaren, Gävle University Authorities Region Gävleborg (MSHV coordinator), Länsstyrelsen Region Dalarna, Other policymakers from the Swedish government or local municipalities OEMs Volvo, Powercell, Nitiu Cellcentric Table 4. Identified actors in MSHV and actors that affect the development of MSHV, classified into actor groups. Some actors are present in more than one group. Renewable energy providers deliver the electricity that is needed in the production of low-carbon hydrogen. Skyborn Renewables is an offshore wind-farm establisher that recognizes hydrogen as a value-adding commodity to their wind turbine farms, ensuring revenues even during periods with very low electricity prices. Dalavind is another renewable energy producer based in Dalarna that is building wind farms, who have recognized an opportunity to create industrial parks for hydrogen production. However, this requires partnering up with hydrogen producers – something that both Skyborn Renewables and Dalavind sees as an attractive possibility in the future. Statkraft is another renewable energy provider, who focus primarily on hydropower, but who aims to produce hydrogen by establishing their own electrolyzers (currently waiting for permit processes). Another use-case that was mentioned in the interviews was that renewable energy providers also could use hydrogen as an energy buffer, e.g. by converting electricity to hydrogen when there is overproduction and converting it back to electricity when there is shortage. 42 Hydrogen producers use different methods for hydrogen production, with water electrolysis being the most common. Actors within MSHV that currently use water electrolysis are Ovako and Linde Gas. Ovako inaugurated their facility in Hofors in September 2023 and is the first company in the world to use low-carbon hydrogen for heating steel. Currently, the hydrogen produced is only used in their own processes, but they have the capacity to produce more hydrogen that could be sold to other users. Ovako also has other sites in the region that could be converted to hydrogen production sites. For example, the company is planning to build another electrolyzer in Smedjebacken. Statkraft is another actor who is currently looking to expand into hydrogen production through water electrolysis, even though their current operation is primarily focused on hydropower. Although Statkraft are formal members of MSHV, they have no current hydrogen project within MSHV. However, they do have a smaller electrolyzer project in Gothenburg, where they aim to gain knowledge about hydrogen production to produce it more broadly in the future. Even though water electrolysis is the most common method for producing low-carbon hydrogen, some actors have innovated alternative technologies to the electrolyzer, e.g. to circumvent the high electricity requirements of the electrolyzers. One of these actors is Plagazi, who has developed a process for producing hydrogen from waste material, which only uses one-fifth to one-eighth of the electricity compared to an electrolyzer. The interviewee from Plagazi claims that the process is also much cheaper than the electrolyzer and also produces heat as a byproduct – which could be sold to other actors. The drawback is that CO2 is also a byproduct of the process, which means that CCS has to be used for the process to become low-carbon. There is currently no commercial scale hydrogen production based on this process, but a proof of concept has been demonstrated successfully. Even though the process is very novel, actors like Alleima – a steel manufacturer and hydrogen user, vouch for its legitimacy and see themselves becoming a possible buyer from them in the future. Storage providers/distributors play a crucial role in driving the hydrogen economy forward by investing in infrastructure including pipelines, storage facilities, and carbon capture possibilities. One actor who is developing technologies for safe and effective storage of hydrogen is Nitiu. Since hydrogen is an explosive gas, safety is a high priority for all use-cases. Therefore, Nitiu has created an innovative storage solution with a durable structure, which they aim to sell to many actors in MSHV. Nordion Energi is an actor who is a transmission system operator (TSO) for the Swedish core grid for natural gas who now has the ambition to build the infrastructure that is required for the hydrogen economy. Today, Nordion Energi is engaged in two major hydrogen infrastructure projects, namely, Nordic Hydrogen Route and Baltic Hydrogen Collector. Both projects have received the labeled “projects of common interest” by the European Commission. Another actor is Hydri, who through strong financial support is building a national hydrogen fueling network (24 fueling stations finished by April 2025). 43 Furthermore, Inlandsbanan is a governmental owned actor who operates a 1000 km long railway that is planned to be used to transport hydrogen produced by Plagazi. Hydrogen users are the actors who have an interest in buying and utilizing hydrogen, either in industrial processes, as fuel, or as energy storage. Hauliers are one actor type that could utilize hydrogen, since they consume high volumes of fuel on the road and need to find alternative fuels to reduce their climate footprint. MaserFrakt is a haulier that operates in the region who has established a hydrogen fueling station and bought their first truck fueled exclusively with hydrogen (MaserFrakt, n.d.). According to the interviewee from MaserFrakt, Hydrogen is one of three main alternatives in the change, besides liquefied biogas and battery electric. Moreover, steel producers could also be a major consumer in a hydrogen economy, since the fossil fuels currently used in their processes represents 10% of Sweden's total carbon dioxide emissions, which could be exchanged for hydrogen (Alpman, 2017). In an attempt to lower their substantial carbon footprint, steel producers are amongst the early adopters of hydrogen today. Besides being a hydrogen producer, Ovako has started to use hydrogen in its steel heating processes, where it replaces fossil fuels. Alleima is another steel producer who aims to be a significant user of hydrogen. They are currently using relatively small amounts of low-carbon hydrogen (135 tons/year) as a shielding gas in their processes, which are produced and transported (in pipelines) from Linde Gas. In the future, they aim to use considerably more hydrogen in their molding processes, where they currently use natural gas. Another actor in the region who aims to use low-carbon hydrogen from Linde Gas is Green Iron, which plans to use hydrogen in the production of their fossil- free sponge iron. Furthermore, Sandvik is another member of MSHV that is using hydrogen in their furnaces for heating and molding. Besides the mentioned hydrogen users, there exists other potential use-cases which no actor currently pursues, including glass production, fertilizer production, and electric grid services, which may become hydrogen users in the future. Grant providers have the purpose of providing financing for investments for the green transition in Sweden, where hydrogen plays a central role. Energimyndigheten mainly finances projects within novel technologies that can reduce carbon dioxide emissions, where the risk for the investor is very high. The funding is provided through the initiative called Industriklivet, which is partly funded by the European Union. Naturvårdsverket sponsors other types of investments related to more proven and mature technologies, e.g. hydrogen fueling stations, pipelines and storage, through an initiative called Klimatklivet. Another grant provider is Tillväxtverket, which also provides financing for actors in MSHV. Incubators and research institutes supports actors in the value chain by providing new knowledge and by creating networks between the actors. The interviewee from Sandbacka Science Park described that they aim to reduce uncertainty, connect actors with each other, and concretize ideas – where 44 hydrogen is a field that they are currently exploring. Furthermore, Gävle Harbor is a logistical center that acts as a facilitator between actors by helping to initiate projects adjacent to the harbor. For this reason, the harbor could be labeled as an incubator in the value chain. Another interviewee, which is a hydrogen researcher at University of Luleå, has among other initiatives been a part of developing Sweden’s national hydrogen strategy (currently not used) and is today responsible for CH2ESS – a hydrogen research network with over 100 researchers. Other incubators are Dalarna Science Park, RISE, University of Gävle, and Mellansvenska Handelskammaren, but they were not interviewed in this study. Authorities have various tasks in driving the transition to a fossil-free society and act on a municipal, regional, or national level. Some public actors in the region, like Region Gävleborg, Länsstyrelsen, and Region Dalarna are directly engaged in MSHV with their geographical presence. A common role that these actors have is being facilitators and trying to bring actors together with the purpose of promoting the hydrogen development in the region. The coordinator of MSHV at Region Gävleborg is for example responsible for holding the conferences. Furthermore, Länsstyrelsen is working with regional strategies within energy and climate. Hydrogen does have very little presence in the current strategy but its role could possibly increase in the upcoming strategy. Regarding public actors on the national level, the Swedish government is the actor who decides the overarching hydrogen strategy, as well as establishes laws and regulations that have a large impact on MSHV. OEMs are actors who are not directly involved in the hydrogen value chain, but who are nonetheless influencing the development of MSHV in significant ways. One OEM actor is Powercell, who play a central role within the transport industry, since they develop and produce fuel cells that are used in vehicles or for stationary applications. Powercell is a member of MSHV, and produces hydrogen fuel cells based on PEM (proton exchange membrane) technology. Additionally, besides selling their fuel cells, they also sell the services to help customers integrate their systems into their operations. Another OEM who can contribute greatly to the hydrogen development is Volvo, who aims to develop and produce trucks running on hydrogen – therefore providing hauliers the possibility to buy trucks that run on hydrogen. Volvo has invested heavily in developing their hydrogen trucks and related technologies, since they identify hydrogen fuel cells as a replacement for diesel engines, in addition to being a complement to electric engines. Volvo sees their future product portfolio as being a mix of hydrogen and electric vehicles, with different types of use-cases (e.g., hydrogen trucks being used for longer distances and where charging is inconvenient). In 2023, Volvo tested their first hydrogen trucks on public roads (Volvo Group, 2023). 45 4.1.2 Networks A network is an interconnection between actors in the TIS. While the largest network in the region is the organized MSHV-platform, the interviews revealed several less organized networks between the actors (see Table 5 or Appendix A for a visual illustration). However, due to the sheer number of actors and possible network constellations, the identified networks are not exhaustive, but rather used to showcase the most important interconnections and collaborations in MSHV. In the following section the identified networks are presented. Actor Groups Network 1 Network 2 Network 3 Network 4 Network 5 Network 6 Network 7 Network 8 Renewable energy providers MSHV Svea Vind Offshore Skyborn Renewables Hydrogen producers Ovako Linde Gas/Plagazi Ovako Storage/Distri bution Nordion Energi Hydrogen users Ovako MaserFrakt Green Iron, Alleima Alleima Volvo Group Grant providers Energimyn -digheten Incubators and Research institutes Gävle Harbor Authorities Muncipalities of Gotaland, Åland and Bornholm OEMs Volvo Group Volvo Group, Cellcentric Support outside MSHV Nel Hydrogen, H2 Green Steel, Hitachi Energy Gasgrid Finland, OX2, Copenhagen Infrastructure Partners Lhyfe, ABB Daimler Trucks Table 5. Networks. 46 Network 1 - MSHV collaboration platform The largest network in MSHV is the collaboration platform of the MSHV initiative itself, which includes over 30 different actors (Region Gävleborg, 2023). This network is facilitated mostly by inviting members to conferences and seminars, which are held one to two times per year. The get- togethers could have the following agenda: (1) an introduction by MSHV representatives from the regional authorities about news, trends and opportunities in the region, (2) presentation from selected actors who introduce their company and their existing projects related to hydrogen, (3) open discussions where a microphone is passed around, and (4) physical tours and showcases of different hydrogen projects. The coordinator of MSHV explains that the purpose of the MSHV get-togethers is to provide a platform for discussion – to showcase collaboration possibilities amongst the members. The purpose is also to share risk among the members by, for instance, sharing investment costs, being transparent with supply and demand, and collectively applying for grants. The coordinator also mentioned that some actors are reluctant to share information about their innovation processes, and by having a platform that ensures mutual trust, more information can be shared between actors in the region. Since its inception in 2021, the MSHV collaboration platform has received great attention from actors, both in terms of new members joining and in participants on the get-togethers. The majority of the interviewed actors had positive feedback about the get-togethers and felt that they were meaningful. Common examples that were brought up were good communication, solid coordination and leadership, and that a balanced mix of policymakers and representatives from industry were present. However, some actors expressed concern that the collaboration resulted in very few practical projects. For example, one actor expressed dissatisfaction that the get-togethers were merely talks with no action, and therefore a waste of time. This frustration was based on several experiences where the actor had approached other members and tried to enter into partnerships, but where the other actors had rejected the proposal. According to the rejected actor, other actors do not try hard enough, since they are reluctant to take financial risks. This resulted in the rejected actor having to enter into a sub-optimal partnership with another actor outside of MSHV much further away. From the perspective of the other actors, the uncertain financial gains from the proposed partnerships was indeed why they refused to collaborate, since this type of partnership would require large investments in infrastructure between the actors. This situation indicates that there is an inconsistency between the expectations of some members with high ambitions, and the actual purpose of MSHV – which is not specifically to initiate projects. This in turn creates a discrepancy in how MSHV is perceived. When discussing the relatively few projects currently taking place in MSHV, the MSHV coordinator explained that the projects should be initiated by the members themselves – the coordinators of MSHV 47 do not force actors to participate in a project. However, the coordinator acknowledges that more actors have to be willing to initiate projects and that more have to think long-term rather than maximizing profit in the short-term. If more actors would participate in projects, the financial risks of these projects would be lower, and the MSHV collaboration platform could also start applying some pressure on the actors who do not contribute. Network 2 - Electrolyzer collaboration Network 2 represents a collaboration that was formed in the creation of one of the most prominent hydrogen projects within the region – a hydrogen production plant in Hofors. Ovako had the central role in the project, where they built an electrolyzer from which they use to produce hydrogen that they use in their steel production. The project required partnership with companies like Volvo, Hitachi Energy, H2 Green Steel and Nel hydrogen, who contributed with different components, technical solutions (Ovako, n.d.), and financial support. The project also received funding from the grant provider Energimyndigheten through the program called Industriklivet (40% of the planned investment cost). These partnerships and financial support were essential in the establishment of the hydrogen plant – which is today the largest electrolyzer in Sweden (20 MW and 3,880 cubic meters of hydrogen per hour) and the first low-carbon hydrogen used in steel heating in the world (Ovako, n.d). Currently, Ovako is also investigating another potential electrolyzer project in Smedjebacken in the southern region of Dalarna. Network 3 - Gävle Harbor Gävle Harbor acts as a logistical center, which is an important facilitator for the hydrogen development in the region. The interviewee from Gävle Harbor described that their role in the hydrogen development in the region is to bring actors together, offering land and assisting in permit processes. Network 3 is a collaboration between several actors connected to Gävle Harbor. Adjacent to the harbor, Svea Vind Offshore plans to establish a wind turbine farm to produce electricity that could be used to produce hydrogen in cooperation with the harbor, but this project has been delayed due to a rejected permit application. The haulier MaserFrakt, who regularly use the harbor to transport goods, has shown great interest in a partnership with Svea Vind Offshore, and have signed an agreement where they will buy hydrogen that will be used in their trucks (Cision, 2022). Network 4 - Baltic Hydrogen Collector The Baltic Hydrogen Collector is a cross-border project to create an offshore hydrogen pipeline infrastructure in the Baltic Sea. Nordion Energi and Gasgrid Finland Oy are initiating the project in collaboration with the wind power developers OX2 and Copenhagen Infrastructure Partners. Furthermore, the municipalities of the islands of Gotland, Åland and Bornholm have signed a letter of intent to become collaborators in the projects, since their beneficial geographical location is making 48 them important nodes in the pipeline. Besides providing the Baltic Sea countries with better and quicker access to green hydrogen, the infrastructure will also enable balancing of the power grid in the region. Moreover, the project was on November 28, 2023, announced to be a “project of common interest” by the European Commission (BHC, n.d.). The Baltic Hydrogen Collector project identifies mid-Sweden as a demand center, and plans to have a node there in the future (Baltic Hydrogen Collector, n.d.) – thereby creating an important infrastructure connection between MSHV and other European countries. Network 5 - SoutH2Port Network 5 is a collaboration between Skyborn Renewables, Lhyfe and ABB that aims to build one of Europe’s largest hydrogen plants in Söderhamn – a project called SoutH2Port. The hydrogen plant will be powered by electricity from Skyborn Renewable’s planned 1 GW Storgrundet offshore wind turbine farm and is expected to produce 240 tons of hydrogen per day, 88.000 tons annually (Lhyfe, 2023). Lhyfe produces systems for low-carbon hydrogen production and will together with Skyborn Renewables operate the plant (Lhyfe, 2023). Furthermore, ABB brings technical expertise for optimizing the integration of hydrogen and electricity production (ABB, 2023). Network 6 – Steel producer collaboration Network 6 is a tenant agreement