Navigating an emerging innovation ecosystem: a case study of fuel cell innovation in Taiwan

ABSTRACT

The Taiwanese government has actively promoted and established policies for developing fuel cell research and industrial activities since the early 2000s. The activities, including those of Taiwanese policy actors, private firms, and research institutions, are part of a fuel cell innovation ecosystem. Through an embedded case study, this paper examines past national policies, sectoral-level developments, and how a firm has navigated the environment. The analysis is based on the notion of an innovation ecosystem, which analytically covers micro, meso, and macro levels and has more porous boundaries than a national innovation system. Hence, the contribution of this paper relates to creating a deeper understanding of what is happening under the surface of government policy and what path firms can take. Thus, it contributes to studies on latecomer strategies in technology sectors of the dynamic interaction between policy, sectoral development, and firm strategy.

KEYWORDS:

• Innovation ecosystem
• Taiwan
• fuel cells
• latecomer
• policy
• firm strategy

1. Introduction

Wind, solar, and water comprise the bulk of clean and renewable energy sources. In Europe, electricity is traded within a pan-European grid to balance energy demand when solar and wind energies are over- or undersupplied. In smaller countries and regions with more extensive needs for self-sufficiency and where cross-continental trading is not an option, alternative solutions are needed to a greater degree for more efficient use and storage of available energy sources (Papaefthymiou & Dragoon, 2016). Such technologies could be, for example, batteries or fuel cells. On an island such as Taiwan, with a more closed energy system, storage solutions such as fuel cells have been seen as an essential technology to leverage clean energy sources (Chou et al., 2019). Moreover, alongside efforts to create a domestic supply base of fuel cells catering to local needs, the Taiwanese government has pushed for developing a fuel cell industry that can be competitive in an international context.

While Taiwanese firms are latecomers and have had to play technological catch-up, the global fuel cell industry still has difficulties scaling up successful technical demonstrations and applications of fuel cell technologies (Chou et al., 2019). A series of causal factors impede their broad use and commercial applications in society. For example, the lack of a sizable market hinders the formation of economies of scale, and large-scale deployment is lacking, which impacts the possibility of verifying the consistency and reliability of the technology (Wang et al., 2018). Nonetheless, fuel cells remain a technology area that garners the interest of the public and private sectors because of their advantage as a stable power source.

The Taiwanese government has actively promulgated policies for developing fuel cell research and industrial activities since the early 2000s. The activities, including those of Taiwanese policy actors, private firms, and research institutions, are part of a fuel cell innovation ecosystem (Ministry of Economic Affairs, 2016). The private sector plays a vital role in such a system and is integral to developing and disseminating fuel cell technology that follows the standards of a global technological regime. On the one hand, given the nascent environment, firms, especially SMEs, need government support to develop their business viability (see Tsai & Liao, 2017). On the other hand, firms cannot rely on government subsidies for their survival in the long term (Shih & Aaboen, 2019). Reliance on government support can lead to strategic conflicts when technology firms need to organise global innovation activities simultaneously as they manage within their local institutional settings (see Hu et al., 2017). This challenge highlights the need to understand the adequacy of innovation and industrial policies in developing the business viability of high-tech firms in emerging technological sectors. This study aims to elucidate this issue by asking the following questions:

• What underlying dynamics shape and impede sectoral innovation in the latecomer setting of the Taiwanese fuel cell sector?

• How has business viability been achieved in this sector?

To answer these questions, the paper uses the notion of an innovation ecosystem, which analytically covers macro- to micro-levels and has more porous boundaries than a national innovation system (see Granstrand & Holgersson, 2020; Russell & Smorodinskaya, 2018). The framework allows for an analysis that can integrate the study of system-level characteristics (e.g. Chou et al., 2019; Xiong et al., 2022) and latecomer firm strategy (e.g. Chen & Zheng, 2023; He & Sun, 2023; Huang & Intarakumnerd, 2019). To this end, this paper uses an embedded case study of the Taiwanese fuel cell sector, examining past national policies, sectoral-level functions and dysfunctions, and how a case firm has navigated in its environment. The contribution of this paper focuses on creating an integrated understanding of what is happening under the surface of innovation ecosystems formed through government policy, particularly the feedback loops between firm actions and national systems.

The lessons from the analysis of Taiwan can inform other small latecomers and export-oriented countries with large SME sectors on how innovation policy can be developed to promote the production and use of fuel cell technologies and clean energy innovation overall. The paper is structured as follows. The next sections present relevant literature and the paper’s methodology. An empirical case description and analysis follow after that. The paper’s final section presents the conclusions.

2. Relevant literature

2.1. Visions of systemic innovation as the primary policy driver and challenges

The systemic aspects of innovation have, for the past 40 years, been an essential topic of both scholarly and policy interest. Innovation has been described through a system perspective on national, regional sectoral, or technological levels (Hekkert & Negro, 2009; Lundvall, 2007). Such perspectives allow policymakers to derive policies for industrial development within the confines of a country or technological sector. Therefore, the notion of an innovation system has heavily influenced innovation policy (Borrás & Edquist, 2013). Public actors play an integral role in stimulating the environment to be conducive to innovation through, for example, providing funding, establishing institutional structures, and communicating good practices (Xiong et al., 2022). At its fundamental level, it refers to ‘the network of institutions in the public and whose activities and interactions initiate, import, modify and diffuse new technologies’ (Freeman, 1987, p. 1). This description has encouraged regional and national governments to focus on structural components, such as the composition of actors and institutions in the innovative environment. However, a critique of basing innovation policy on the construct of an innovation system is that it is challenging to evaluate and empirically validate the actual efficiency of the system, especially in emerging sectors (Lema et al., 2020).

Scholars have argued that innovation system approaches cannot detect idiosyncrasies within the system at the level of individual firms and networks (e.g. Hung & Whittington, 2011; Markard & Truffer, 2008). Markard and Truffer (2008) specifically state that the lack of focus on understanding individual firm behavior and their interactions with other actors within the system limits the understanding of innovation processes. The lack of individual firm behavior is a weakness because there are considerable variations in innovation processes and paths within a system (Huang & Intarakumnerd, 2019; Ortiz, 2013). Hence, a one-size-fits-all configuration that does not allow for internal flexibility might lead to less innovation. Hung and Whittington (2011) argue that prearranged institutional directives can limit innovation and require firms to follow institutionally embedded practices. Practices that have become institutionalised include, for example, funding agencies that require innovators to interact with certain actors within the innovation system and in particular ways (such as forming a value chain). Moreover, Ortiz (2013) states that firms should follow specific stepwise requests from supportive actors to garner legitimacy. Hence, an innovation policy based on innovation system perspectives can be too conformal and constrain firms in their innovative activities by not allowing for divergences in firms’ innovation strategies (Hung & Whittington, 2011).

Analyzing substructures or agency within systems might be helpful, especially in contexts characterised by high uncertainty (Dattée et al., 2018). Focusing on new firms that emerge within the environment is of particular interest and value. The rationale is that these actors might contribute the greatest to value creation but also need the most assistance in the system to innovate (Clarysse & Bruneel, 2007).

In summary, prescribed innovation practices might not be the most conducive for innovation but result from a streamlining process by policymakers to fit policies with the idea of systemic innovation and the belief in the possibility of coordination (Borrás & Edquist, 2013). Consequently, the very idea of systemic innovation can inhibit innovation in several instances, as argued above. In particular, the inclusion of different actors into an environment without understanding their contextual roles and interaction patterns diminishes its practical value and requires better understanding (see Hekkert et al., 2007). Here, the firm’s management capabilities are integral to understanding transformative processes (Dattée et al., 2018).

2.2. Innovation ecosystems and network dynamics

Given that considerable public resources are allocated by policymakers to develop regional and national innovation support structures and to fund various system actors, it is essential to critically explore the relevance of the supportive environment and how it fosters diverse innovators (Mathews et al., 2011; Shih & Aaboen, 2019). The need for evaluation relates for example to the engineering feasibility of technologies and their commercial viability (Wang et al., 2018). Therefore, researchers have increasingly emphasised the need to explain the dynamics of business networks and innovation ecosystems.

The notion of innovation ecosystems can provide a context to describe the dynamic interactions occurring within networks and their relationship to institutional contexts. While several conceptualisations of innovation ecosystems exist, Granstrand and Holgersson (2020) describe that definitions often emphasise collaboration and actors. As such, inter-organisational interactions within an ecosystem become a focal point of interest (Still et al., 2014). Rubens et al. (2011, p. 1737) emphasise that ‘synergies in components of the innovation ecosystem are observed through network analysis’. These networks work across the micro, meso, and macro levels (Russell & Smorodinskaya, 2018).

Ecosystems are adaptive systems (Martin & Sunley, 2007). The boundaries are more porous than in the innovation system framework, which is a closed system (see Russell & Smorodinskaya, 2018). In the innovation system, networks interact through incentives driven by institutional frameworks (Bergek et al., 2008). Collaborative dynamics are of interest from the innovation ecosystem perspective (Russell & Smorodinskaya, 2018). Hence, networks play an essential role as integrators of resources and opportunities. Crossing macro, meso, and micro levels, Granstrand and Holgersson (2020) identified three main components of innovation ecosystems that explain innovation dynamics, including actors, activities, and artifacts. Actors include firms, government actors, universities, and research institutes (Laage-Hellman et al., 2020). Firms can take on different roles in a network, ranging from suppliers to buyers, as competitors or partners. Moreover, the network investigation covers a broad set of actors, including some that government actors might not identify as part of a national or sectoral system. Hence, parts of an innovation network might be politically connected while others are not (Linné & Shih, 2013). Activities between actors are undertaken within network structures. Such activities include those related to developing, producing, and using technologies (Shih & Linné, 2016). The innovation ecosystem thus comprises ‘a network of interdependent actors who combine specialized yet complementary resources and/or capabilities in seeking to (a) co-create and deliver an overarching value proposition to end users, and (b) appropriate the gains received in the process’ (Walrave et al., 2018, p. 104).

As diverse actors contribute to varied and idiosyncratic needs for knowledge and competencies in realising innovation, several studies have suggested going beneath the system level to analyze innovation in a system or environment (Chen & Zheng, 2023; He & Sun, 2023; Huang & Intarakumnerd, 2019; Hung & Whittington, 2011; Waluszewski et al., 2009). This paper investigates specific inter-organisational network dynamics, embedded within innovation ecosystems, which can reinforce understanding of how institutional support structures help or not. This study also takes a firm’s perspective from the fuel cell sector in Taiwan to identify the dynamic features in the interaction between the various actors in emerging sectors in the latecomer context. The firm builds relationships that form network structures embedded within the system. Using a firm as a probe means that the actors in a fuel cell innovation network can be identified by looking at the interactions of a focal firm. Moreover, these actors’ interaction patterns and roles can be studied throughout the innovation process. The focus on interaction enables an analysis of the system and its internal dynamics, providing insights for policy development. Figure 1 illustrates the analytical framework.

Figure 1. Innovation ecosystem and network dynamics.

The analysis starts with a national-level analysis of policies promulgated to stimulate innovation and entrepreneurial and industrial activities. Following the national-level analysis, a meso level analysis of the impact of these policies on a sectoral level follows. To gain a deeper understanding of the underlying dynamics, an analysis of a firm’s network activities shows the interactions at the micro level. The firm’s interactions show feedback loops occurring throughout the macro, meso, and micro levels.

3. Methodology

3.1. Research design

Case study methodology is widely used in qualitative research because of its practicality in illustrating complex, in-depth phenomena (Eisenhardt, 1989). This study describes fuel cell innovation from an embedded case study perspective in Taiwan. An embedded case study approach involves more than one subunit of analysis (Yin, 2009). The rationale for an embedded case study with multiple subunits is to allow for a ‘broader exploration of research questions’ (Eisenhardt & Graebner, 2007, p. 27). The starting and ending points are the understanding of the case as a whole; however, during the analysis, the perspectives of different units of analysis are taken (Scholz & Tietje, 2002). Using case study methodology allows for analytical generalisability, which compares case findings to existing literature (Yin, 2009). As Stake (1995, p. 8) notes, the purpose of case study research is ‘particularization, not statistical generalization’.

This paper focuses on macro, meso, and micro level developments (using a framework inspired by Granstrand & Holgersson, 2020; Russell & Smorodinskaya, 2018). The embedded case study allows for analyzing the interconnections and dynamics between the different levels through network analysis. More specifically, the empirical study departs from a general description of the national policies that have served as a foundation for promoting the development of the fuel cell sector in Taiwan. An examination of developments at the sectoral level follows the overview of policies. Thereafter, we follow a fuel cell firm and how it navigates within its network (micro), crossing both the macro and meso levels. The subunits of analysis were selected because they enable a deeper understanding of the empirical phenomenon and their connection to theoretical constructs (Eisenhardt & Graebner, 2007). We chose the case firm M-Field because of its long history in the Taiwanese fuel cell sector stretching from 2000 to 2020. The firm’s activities also cover domestic and international spheres. Thus, the case study firm was valuable for understanding system and network dynamics. This research design follows the logic of Markard and Truffer (2008), who argue for including micro-level analyses to understand system-level challenges.

3.2. Data collection

We collected data in Taiwan between 2016 and 2020. One of the co-authors of this study has been working in the fuel cell industry for over 14 years, including at one of the major fuel cell research institutes, the Taiwan Institute of Economic Research (TIER). His background enabled the access to data and a deepened understanding of the sector. Two methods were employed to collect first-hand data: participant observation and in-depth interviews. These methods allowed us to authenticate data generation and produce a more in-depth understanding of the empirical phenomenon (Macdonald, 2012).

We conducted 31 formal interviews (see Table 1 below) with executive managers of fuel cell firms, public institutes, and universities from 2016 to 2020. These interviewees were either referred to the authors by specialists in public institutes or recommended by participants in various meetings and workshops organised by important fuel cell stakeholders. Each interview lasted 30–180 min and we interviewed some participants several times. The interviews provided information about the respondents and their organisations’ activities and roles in the fuel cell sector, of (at the time) present and past events. The interviews covered the strategies the organisations were using and the discussion of various important policy and corporate events. We did not record the interviews but took detailed notes. If the data gathered in the initial interviews was unclear, the interviewees were contacted again by phone or other appropriate means to obtain clarification. Moreover, the senior manager of TIER and chairpersons of firms 2 and 3 were interviewed several times to cross-check for consistency and obtain further verification.

Table 1. List of interviewees, 2016–2020.

Respondent	Affiliation	Job position
1	M-Field	Senior manager
2	Firm 1	Chairman
3	Firm 2	Chairman
4	Firm 1	General manager
5	M-Field	Senior manager
6	University 1, fuel cell R&D	Professor
7	Public Institute 1, Fuel cell research	Director
8	University 1, fuel cell R&D	Researcher
9	M-Field	Senior manager
10	Firm 2	Chairman
11	Taiwan Institute of Economic Research (TIER), fuel cell research	General Manager
12	Firm 3	General Manager
13	Firm 2	Manager
14	TIER, fuel cell research	Director
15	Public Institute 1, green energy research	Chief engineer
16	M-Field	Senior Manager
17	TIER, fuel cell research	General Manager
18	Central government, Renewable energy division	Director
19	Local government, renewable energy promotion division	Deputy Director
20	TIER, fuel cell research	Researcher
21	University 2, green energy research	Professor
22	Local government, renewable energy promotion division	Manager
23	Firm 3	Chairman
24	M-Field	Senior manager
25	Firm 2	Chairman
26	Firm 4	Vice President
27	Public Institute 1, green energy research	Manager
28	Firm 3	Chairman
29	M-Field	Senior manager
30	TIER, fuel cell research	General Manager
31	M-Field	Senior manager

Moreover, we conducted informal meetings and had dialogs with key people in the Taiwanese fuel cell sector to increase our understanding of the industry. These meetings were with business managers, industry leaders, and university professors. The information was additionally corroborated through a triangulation approach (Yin, 2009) using secondary sources such as policy reports, newspaper articles, and company reports. As interviews are often recollections of events, triangulation became an important method to increase the trustworthiness of the case study. For the empirical study, we used sources such as government websites, news items from DigiTimes (one of the largest technological development news platforms in Taiwan), the Taiwan Hydrogen and Fuel Cell Partnership (THFCP, established in 2002), the Industrial Technology Information Service (ITIS at Industrial Technology Research Institute), and the Taiwan Association for Hydrogen Energy and Fuel Cell (THEFC). In addition, we extracted information from company reports and public statements to review the firms in Taiwan.

3.3. Data analysis

The development of the case and subsequent analysis were performed abductively (Dubois & Gadde, 2002). The innovation ecosystem concept helped us structure the case study by highlighting the system perspective and understanding dynamics across macro, meso, and micro levels (Russell & Smorodinskaya, 2018). Rubens et al.’s (2011) suggestion that network analysis could show these dynamics across levels was important for the decision to analyze how firms navigate within a particular innovation ecosystem.

By collecting data on general policy and sectoral developments, the paper illustrates the macro and meso levels. The case study started with Taiwanese policy initiatives from the early 2000s until 2020. We identified the major industrial events in the Taiwanese fuel cell sector during the period and investigated how various system resources were combined. The analysis of government policies showed prescriptions of resource use and actor interactions. However, to understand deeper system dynamics, such as how firms were inhibited or supported, we also needed to look beyond the availability and provision of resources and policies and investigate how firms interacted. Here, the case of M-Field helped illustrate micro-level interactions. The embedded case study allowed for the analysis of how the innovation ecosystem has contributed to firm-level innovation activities, albeit indirectly, where firms are the drivers. The research design allowed for a deeper explanation of innovation ecosystem dynamics and identifies various benefits with policy support, dysfunctions, and examples of inertia.

4. Case description

4.1. Macrolevel developments: government policy supporting fuel cell research and business in Taiwan

Although fuel cells have existed for over 200 years, commercial applications were not available on a larger scale until the 1990s, primarily in the United States and Japan. In 2001, the Taiwanese government developed research and innovation policies for fuel cell technology. The first policy came after the Sixth National Science and Technology Conference, the main forum for shaping science and technology policy roadmaps in Taiwan organised every four years. Policymakers believed Taiwan could have an edge as a latecomer in a relatively new commercial technology area by relying on a solid manufacturing base and policy guidance.

Government actors developed incentives for universities and private firms, and funding was allocated to public research institutions to start initial activities. Policy guidance focused on portable electronics, electric scooters, and small stationary power generators. The government provided grants to universities through the Ministry of Science and Technology (MOST) to direct basic research activities. The Ministry of Economic Affairs (MOEA) allocated grants to businesses to work on fuel cell innovation and product applications and development. The director of the renewable energy division at the Ministry mentioned:

Taiwan is used to being a latecomer, and the government strategically leads its industrial development. However, our resources are limited … so public funding cannot be large. Development has to start from small demonstration sites to gradually gain credit and legitimacy. We know this is difficult for small firms. However, we also see that it reinforces their competitive advantages and capabilities if they can survive through the early stage.

A new legislation, the Renewable Energy Development Act (REDA), was drafted in 2002 to enable renewable energy development, growth, and installation. The bill identified fuel cell technology as an essential technology, but it took seven years to pass it as legislation. In 2007, Taiwan’s Bureau of Energy (BoE) included in the national Energy White Paper a goal to advance the development of fuel cell technology. To that end, the Taiwanese government established policies in 2009 for early market demonstrations and verification of fuel cell technologies. In 2009, another government proposal, ‘Green Energy Rising’ set the ambition to turn Taiwan into a global fuel cell manufacturing location. While the Energy White Paper provided strategies and targets, the Green Energy Rising proposal developed an action plan and fashioned incentives for R&D actors to start projects within fuel cell technologies.

4.2. Sectoral level dynamics: Taiwan’s firms looking for niche opportunities

With government incentives, some Taiwanese firms quickly identified market opportunities. For example, Delta Electronics, a global power supply firm, and Tatung, a leading consumer appliances manufacturer, began working with overseas fuel cell firms. Delta Electronics established a fuel cell business unit by engaging with Ballard, a global fuel cell system manufacturer.¹ The activities of Tatung and Delta Electronics in the research and commercialisation of fuel cells were limited as the companies deemed the domestic market too small. However, with government funding, smaller side projects could be undertaken. The chairman of one of the new firms explained:

We believe that the fuel cell has great potential and will become one of the driving technologies in the future, especially in the green and sustainability sector  … we are waiting for the potential arrival and striving to acquire government funding, even just a small demonstration site … we try to stay flexible for our strategy and happy to collaborate with any domestic and international stakeholders if there is an opportunity for fuel cell applications.

Ballard and Hydrogenics were pioneers in the North American commercial fuel cell sector and had accumulated a considerable technology portfolio. They used their advantage to sell technology and key components to Taiwanese firms and research institutes. Ballard and Hydrogenics also assisted Taiwanese firms in creating different product applications in stationary power, ships, scooters, and mobility. The Taiwanese policy push further encouraged smaller firms to enter the industry and contributed to the forming of a number of start-up firms.

Several local firms responded to the call for fuel cell demonstration and verification projects and applied for government subsidies. Taiwanese firms demonstrated a total of 231 systems between 2009 and 2013. Many new and smaller firms competed for government funding with similar proposals. Among these projects, one-third related to stationary power applications. Two-thirds were related to mobile units such as scooters, small vehicles, forklifts, and ships. Some firms were more successful than others, demonstrating the ability to develop proposals in line with government requirements and technological competencies. For example, most mobile demonstrations came from a firm established in 2000, Asia Pacific Fuel Cell Technology (APFCT). The MOEA devised in 2014 a new plan to create viable firms. The plan was to encourage firms with different core capabilities to join alliances and apply for projects. Many start-up firms with limited capability to fund their activities were forced to exit after the new policy.

The support for demonstration and verification projects was insufficient for most local firms to thrive. After finishing the four-year demonstration and verification phase valued at 2.38 billion New Taiwan Dollars (NTD) (approximately 83.6 million US dollars), Taiwanese fuel cell actors had gained knowledge and experience, but the environment was still not ready to enter the next stage where policymakers planned to create a manufacturing base and increase investment in viable technologies. Furthermore, to be commercially feasible, the robustness of technologies needed to be further verified on a larger scale. For scaling up larger markets and international partnership were needed. Incumbents such as Tatung and Delta Electronics saw business opportunities because of the government push, but they were reluctant to increase efforts because of the immature national market. With their more robust resource endowments, they could have supported the sector by connecting existing value chains and industrial networks, but the perceived business risk was deemed too great. The reluctance of incumbents left much of the development of commercial activities in the hands of newly established firms. The chairman of one of the interviewed firms explained the situation:

The big companies are too busy to take good care of immature fuel cell technologies … they think they can buy us when the market is ready. However, we are not for sale, at least not easily if I am still here  … we now have good business and technology collaborations with some key foreign companies. Our overseas experience and accumulated capabilities open a great window for government funding, although the government always has troublesome requirements and regulations.

4.3. Micro-level developments: case of a small Taiwanese fuel cell firm: M-Field

In this section, we present the case of M-Field, one of the new firms established in the wake of government policies. Figure 2 shows M-Field’s network with national and international actors.

Figure 2. M-Field network.

4.4 Origins of M-Field

M-Field was founded in 2009, but its origins stem from 2001 when Jemmytex, a manufacturer of diesel generators, established a fuel cell project. Although Jemmeytex had no experience with fuel cells, the firm’s management became interested in fuel cell technology after advice from the then president of Yuan Ze University and the Taiwanese government’s formulation of a national direction. Yuan Ze University had established Taiwan’s first fuel cell research center in 2000. The university assisted Jemmytex with fuel cell research by combining research from its lab with resources from the company. Jemmytex also sourced critical components for fuel cells from Ballard and Hydrogenics. The first ten years were primarily a period of learning and development for Jemmytex.

As the government increased its support for Taiwan’s emerging fuel cell sector, the firm received several rounds of funding from the MOEA to fund its research and product development. For example, Taiwan’s ‘Fuel Cell Demonstration and Verification Subsidy Program’ proposed by the MOEA, involved an investment of NT$375 million, with an additional investment of NT$450 million from private companies. Between 2009 and 2014, 23 Taiwanese companies participated in 48 demonstration projects across 14 counties and cities, deploying 231 fuel cell systems with a total capacity of 712 kW (Ministry of Economic Affairs, n.d.). These achievements further induced the ‘Hydrogen Fuel Cell Demonstration Program’ with a total investment of NT$350 million between 2010 and 2013. Moreover, the Taiwanese government awarded grants to several small firms, such as Jemmytex, from the Small Business Innovation and Research Fund for fuel cell system projects between 2003 and 2008. Gradually, Jemmytex’s fuel unit became more competent and gradually upgraded its technological capabilities.

In 2009, the fuel cell technology unit and the engineering team at Jemmytex were spun off and became M-Field, with Jemmytex as the majority owner. The Jemmytex management now viewed fuel cells as a viable business, and the spin-off was to focus solely on such technologies. The M-Field management team mainly saw opportunities with larger demonstration and verification projects and wanted to focus on stationary applications. Shortly after, M-Field launched its first fuel cell generator through collaboration with Ballard and Hydrogenics. The generator was intended as a backup energy source for telecommunication equipment. Concurrently, Taiwanese firms such as Delta Electronics and Tatung proposed similar stationary demonstration projects for use in telecommunication systems and hospital settings with varied power output specifications. Hence, several Taiwanese fuel cell sector actors were ready to test their technologies in settings with commercial users.

4.5. Demonstrations of M-Field’s technology

In 2010, M-Field was awarded a larger hydrogen fuel cell demonstration project in Taipei that aimed to accumulate 2000hours of operation. With this project, M-Field became the first company in Taiwan to accomplish a large system demonstration. Moreover, the firm lobbied the government to subsidise fuel cells as backup power systems for telecommunication systems. With 40,000 telecommunications operating sites in Taiwan, M-Field saw a niche opportunity. M-Field’s demonstration project led the BoE to fund larger fuel cell initiatives. Between 2010 and 2012, M-Field received more grants from the BoE to run demonstration projects that deployed fuel cells to back up telecommunication systems. Two of Taiwan’s largest cell phone service providers operated the telecommunication systems, Chung Hua Telecom and Taiwan Mobile Company.

However, obtaining approvals from Taiwanese authorities was a time-consuming and bureaucratic process. There were several reasons. At the time, fuel cell technology was new to users, and many potential users were concerned about hydrogen safety. Another challenge was that M-Field needed to source fuel cell stack components from Taiwanese local firms as part of the agreement for funding. The BoE implemented this requirement to help local firms and suppliers. However, M-Field wanted to work with companies such as Ballard and Hydrogenics, which could provide reliable technology that was more advanced and of higher quality than those produced in Taiwan. The Taiwanese government had nonetheless set a goal to manufacture key fuel cell components in Taiwan. The consideration of national interest was a common practice in Taiwanese government policy, and the project reviewing committees were very likely to request firms to raise their domestic component ratio in their product as much as possible. The requirement was conditional on receiving a government subsidy.

4.6. International expansion

While most of the activities related to fuel cell technology in Taiwan focused on domestic demonstration and verification, M-Field was starting to target international markets. The problem that the firm faced in Taiwan was an immature market and a nascent regulatory system. With domestic market risks, M-Field sought opportunities in international markets and aggressively established partnerships in the United States, Canada, Italy, Australia, and the United Kingdom. Between 2010 and 2015, M-field expanded its international activities. For example, M-Field participated in fuel cell application trials with Ericsson in Sweden and Telus in Canada. M-Field prototyped fuel cell and water electrolysis with Telstra in Australia and developed fuel cells for forklifts with a well-known manufacturer in the United States. M-Field actively participated in exhibitions in Germany and Japan to showcase its technology and capabilities. A senior manager at M-Field explained the goals: ‘The Taiwan market is too small  … Our strategy is to become one of the critical partners for the international players so that we can grow along and enter international, diverse markets with them’. The strategy was to work with major international players to gain broader recognition and experience. Jemmytex, who believed fuel cells could replace diesel generators, primarily funded the activities. Jemmytex’s support allowed M-Field to circumvent some of the Taiwanese government’s funding conditions. However, some of the funding came from demonstration projects and grants from the Taiwanese government. By 2015, M-Field had increased its registered capital to 180 million NTD (approximately 6.3 million US dollars), compared with 5 million NTD in 2009.

Global activities pressured M-Field to increase its quality standards and be more competitive. M-Field was now in direct competition with global fuel cell system integrators, like Dantherm Power and Idatech, who all used the same stack components from Ballard. M-Field had to apply for internationally recognised fuel cell safety certification to meet the fierce competition and international standards. The international competition was a new situation for the firm as it had previously worked in a domestic Taiwanese market. Focusing on international certification took a year and required considerable effort, M-Field gained product safety certification in 2011 and brought the experience back to Taiwan. M-Field shared the findings with different actors, and one of the most important was the National Bureau of Standards, Metrology, and Inspection (BSMI). The firm’s ambition was to help set standards in Taiwan and establish itself as a key local player. In the Green Energy Industry Rising proposal, the BSMI was responsible for setting norms and standards for fuel cell products in Taiwan. Since M-Field had experience with international standardisation organisations, the firm worked with BSMI to build testing facilities, and together they invited international certification experts to Taiwan.

While M-Field has initiated several international projects and shown potential in the demonstration and development stages, none of these projects have become commercialised. However, the projects increased M-Field’s legitimacy and experience in domestic and international markets. M-Field signed a larger purchase agreement with Ballard in early 2014 to secure the supply of key components and lower costs. Furthermore, M-Field and Danish Dantherm Power, owned by Ballard, agreed to collaborate to reduce competition and create synergies. The idea behind the partnership was to combine the two firms’ product and technology development capabilities and leverage Taiwan’s strong manufacturing know-how to accelerate the development of fuel cell power modules. This collaboration was also believed essential by the partners to support Taiwan’s fuel cell sector. Since 2015 Taiwan has been one of the significant markets for Ballard.

Moreover, Chung-Hsin Electric and Machinery Co. (Taiwan’s current largest fuel cell company) acquired Ballard’s methanol telecom backup-power business in 2016. The company combined its electric power business with fuel cells to create microgrid energy services (Ballard, 2016). The applications of Chung-Hsin were intended for mobile units such as forklifts and trucks. M-Field invested USD 1 million in licensing technology from Ballard. The investment further helped M-Field enter the European market and set up a European subsidiary in Denmark. M-Field and Dantherm Power matched well because they utilised similar stack technology provided by Ballard. M-Field wanted to gain market leverage by establishing a European entity and sending engineers from Taiwan to Europe. Since 2015, the international market has continued to grow, which has benefited M-Field and its partners and helped other Taiwanese players gain better knowledge of technological standards and market opportunities.

5. Discussion and conclusions

This paper seeks to answer the following questions: What underlying dynamics shape and impede sectoral innovation in the latecomer setting of the Taiwanese fuel cell sector? How has business viability been achieved in this sector? To answer these questions, the empirical study first described macro level policies in the national arena that have driven the formation of a local fuel cell sector. As Xiong et al. (2022) assert, government involvement has always been a significant factor for latecomer success by creating institutional, technological, and commercial windows of opportunity.

Our study has described how the Taiwanese government established institutional opportunities for technological development and provided initial commercial openings. These policy activities aimed to create a national system of innovation in fuel cells that could be embedded in a larger global value chain. The modus operandi has been a common policy strategy in Taiwan due to the country’s strong manufacturing capabilities (Chou et al., 2019). For example, Zhang et al. (2010) described the successful latecomer strategy of Taiwanese crystal display panel manufacturers as aligning with existing technological regimes and focusing on incremental innovation. There were several challenges to this national and sectoral level strategy in the Taiwanese fuel cell sector. At the meso level, the empirical case shows that many firms were encouraged to enter the sector because of the availability of government funding. Government funding, therefore, impacted firm formation. Several start-up firms also explored business potential and growth through innovation activities because of the availability of government funding. However, some actors depended on government funding as a business model and could not become economically viable. It gave these actors legitimacy in the national system and became a viable approach to secure further investment and innovation support from policy actors. Nevertheless, some private actors had to divest or scale down without public resources. The findings are in line with Shih and Aaboen (2019), which have described the problems with dependency on government support and note that government policies and support play a role in functioning as catalysts, but practices toward economic viability need to be created by firms themselves. In short, new technology-based firms must create viable products on a competitive market. The focus on gaining government support could work against commercial goals in competitive markets. The examples show the limits of public policy, as prior actors influence new actors in the system (see also Kieft et al., 2020).

This paper’s contribution lies primarily in describing micro-dynamics and their interactions with the macro and meso aspects of innovation ecosystems. Scholars studying latecomer innovation have recently addressed this issue. Although latecomer firm strategies have been discussed (e.g. Chen & Zheng, 2023), micro-level strategies and their interactions with ecosystem dynamics have recently become the focus. For example, He and Sun (2023) identified the importance of relationship and value governance in the innovation ecosystem for latecomer firms. Liu et al. (2023) also raise the importance of connecting macro- and micro-level interactions to understand latecomer innovation. Our study contributes to extant literature (Chen & Zheng, 2023; Chou et al., 2019; He & Sun, 2023; Liu et al., 2023) by deepening the understanding of the dynamic interactions that occur in business networks and the connection to policy governance in the innovation ecosystem. The dysfunctions in the national system do not necessarily indicate that the emerging sector is a failure. They primarily suggest the need for system-level changes and promoting the needs of firms advancing technologically and business-wise. Here, the network interactions in the case of M-Field demonstrate some interesting findings. For instance, although M-Field was partly dependent on government funding and a supportive local environment, a nascent user market and too much replication of strategies in a small pond by local actors meant that other opportunities external to the local environment were needed. Firms like M-Field chose to seek external opportunities and mobilise their resources to engage in international activities when the local environment was not commercially viable. M-Field needed to be up to par with international standards and did so by increasing its exposure to global fuel cell competitors. M-Field could support standardisation efforts in Taiwan by leveraging its international network to be technologically relevant and strengthen innovative capacity through cocreation with foreign actors. Such activities could also reinforce innovation in the Taiwanese fuel cell sector. This case shows the importance of feedback loops in innovation ecosystems (as also argued by Dattée et al., 2018). Firms in small and open countries such as Taiwan must utilise an international network to learn and gain market opportunities.

5.1. Policy and practical implications

Our study acknowledges the need for the government to subsidise technology development and markets in a latecomer setting. Extant literature on the development of new industries in the latecomer setting has described these needs (e.g. Chou et al., 2019; Hu et al., 2017; Pegels & Altenburg, 2020). Our analysis identified opportunities and challenges in the national fuel cell sector and the limitations of government policies supporting its development.

The paper notes that new high-tech firms usually lack resources and established business relationships. Hence, the firms actively navigate the local environment to accumulate resources, capabilities, and business relationships as fast as possible (Aaboen et al., 2016; Laage-Hellman et al., 2020; Shih & Aaboen, 2019). The importance of latecomer policy is the ability of governments to adopt proactive strategies in this area quickly (He & Sun, 2023; Hu et al., 2017; Liu et al., 2023). Policies that focus on forming technology capabilities are especially integral to driving the development of innovative activities. Creating an environment for firms to learn and co-create is essential to policy plans. However, a complex area to develop is the development of business relationships.

Creating viable firms with technological capabilities and sound business models requires exposing them to external competition (see Shih & Aaboen, 2019), as shown in the paper’s empirical study. Business model viability reinforces competitiveness in international markets and helps identify commercially less viable firms relying on government funding. An important implication is that policy actors must establish barriers for firms relying too much on government subsidies. The balance can be challenging to find, as the government needs to recognise the need for accumulating knowledge producers and co-locating actors involved in technology development and commercialisation. Therefore, a pragmatic policy approach that not only reinforces innovation support but also stimulates entrepreneurial firms to flexibly navigate opportunities for learning, adaptation, and growth in a local setting as well as in an international context is essential (Dattée et al., 2018; Russell & Smorodinskaya, 2018). A pragmatic approach resonates with He and Sun (2023), who argue for developing a latecomer firm’s unique business innovation ecosystem. Huang and Intarakumnerd (2019) have likewise described alternative technological learning paths in the Taiwanese context, which suggests the need to understand how system-level institutions create disparate innovation models.

As standards and technological development cross boundaries, it is essential to integrate the international dimension into policy plans early on, which has also been suggested by Lema et al. (2020). As seen in this study, the use aspects and embedding of technology in various contexts are best done by research or private actors. Firms are especially important, and emphasising learning about international standards and markets is a necessary part of latecomer industrial policy. The latter is essential for gaining deep knowledge of standards and technologies. These findings are relevant for technology innovation in Taiwan and other latecomer economies (for example in Latin America or Asia) that are export-oriented, SME-centric, and involved in technological catch-up.

In conclusion, our study suggests policymakers should have flexible but robust appraisal systems to identify and evaluate whether actors can innovate and be economically viable. Appraisal methods used by policymakers must be periodically evaluated so that firms are encouraged to rely on something other than state funding. Policy action should also focus on stimulating firms to become more competitive at embedding in established business networks locally and internationally.

5.2. Further research

Industrial development requires government intervention in the latecomer setting. However, the efficiency and relevance of these strategies need to be better understood from the micro level of firm interactions. The dynamic processes of micro-meso-macro interactions could be performed using frameworks such as the ones proposed by innovation ecosystems. This paper has used an embedded case study method. To further understand how network dynamics impact sectoral and national level innovation, additional case studies from other industries or countries could be of interest to understand variations within innovation ecosystems or innovation systems.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

Tommy Shih

Tommy Shih is an associate professor in Business administration at Lund University in Sweden. His research focuses on geopolitics, research and innovation policy, and business networks.

Mei-Chih Hu

Mei-Chih Hu is a professor in Technology management at National Tsing Hua University in Taiwan. Her research focuses on latecomer innovation, and technology management.

Justin Che-Ping Chou

Justin Che-Ping Chou has a PhD in Technology management from National Tsing Hua University in Taiwan. He has worked in the fuel cell industry for over 14 years.

Notes

1 Ballard is a leading global fuel cell innovator. According to its annual reports (2021 and 2012), sales were $104 million and gross margin of $21 million in 2020; for 2011, respective numbers were $56 and 7 million. Since 2015, Taiwan has been a major market for Ballard.