Open innovation at the interorganizational network level

Open innovation at the interorganizational network level – Stretching practices to face technological discontinuities in the semiconductor industry Jörg Sydow1 & Gordon Müll-Seitz2

Abstract Previous studies on open innovation have primarily adopted the perspective of single firms. Against this background we explore how open innovation is managed at the network level for the case of the International Technology Roadmap for Semiconductors (ITRS). Using a longitudinal case study, we report on how the ITRS deals with a technological discontinuity affecting the whole industry, namely the pursuit of an existing technological paradigm using a well-established collaborative practice set against a novel and unknown future technological paradigm to be opened up in parallel, even though the necessary knowledge about the ensuing disruption is lacking. This challenge is characterized not only by extreme technological uncertainty, but also by partner-related and procedural uncertainty concerning the key activity of ITRS, namely the setting of future technological milestones (roadmapping). We discuss attempts by the ITRS network to overcome these uncertainties and manage technological discontinuities by extending the well-established practice of roadmapping to an unknown technological paradigm. We introduce the concept of ‘stretching practice’ to highlight this phenomenon, and we contribute to the debate on the open innovation and technological discontinuities that arise when venturing beyond firm-centric approaches.

Keywords:

Technological

discontinuities;

open

innovation;

stretching

practices;

interorganizational networks; roadmapping; structuration theory; semiconductor industry

Accepted for publication in: Technological Forcasting and Social Change 2018

1

Freie Universität Berlin, Germany: [email protected]

2

Techische Universtität Kaiserlautern, Germany. 1

1. Introduction Ever since Chesbrough (2003) introduced the notion of open innovation, collaboration with external partners through the management of in- and outbound flows of knowledge to foster intraorganizational innovation and generate market opportunities has topped the strategic agenda of many firms (Chesbrough and Crowther, 2006; Gassmann et al., 2010). One of the main reasons for the widespread interest in these ideas is that the vast majority of publications stress the many benefits of open innovation, highlighting the opportunity to access ideas external to an organization and to reduce development costs and risks (Chesbrough, 2003; Lichtenthaler, 2011). A wide range of such external ideas and collaborators has been analyzed, including users (lead users in particular), suppliers, venture capitalists and even competitors (e.g., Gassmann and Reepmeyer, 2005; von Hippel, 1986), making the concept applicable to a variety of organizations and sectors. Through these findings, previous research on open innovation has advanced our understanding of why and how organizations engage with external partners when innovating their products or processes. Given the importance of collaboration to individual firms, it is somewhat surprising that interorganizational networks (or networks for short), in which three or more organizations undertake joint activities and constantly pursue realigned objectives (Powell et al., 1996; Provan et al., 2007; Sydow et al., 2016), are seldom considered in the context of open innovation (for exceptions see Necoechea-Mondragón et al., 2017; Powell et al., 1996; van der Valk et al., 2011). After all, networks are very common in innovation management (Freeman, 1991; Gassmann et al., 2010; Maula et al., 2006; Vanhaverbeke, 2006; Vanhaverbeke and Cloodt, 2006; West et al., 2006) and are even considered the locus of innovation in some industries (Powell et al., 1996). If networks are analyzed at all from a perspective of open innovation, the focus is typically on how a single organization can benefit from engaging in a network (Chesbrough and Rosenbloom, 2002; Dittrich and Duysters, 2

2007; Enkel, 2010; Maula et al., 2006). However, such a firm-centric perspective does not fully account for the fact that networks, including innovation networks, may themselves constitute a form of governance akin to that found in a single entity. A similar firm-centric approach is present in studies of technological discontinuities (Tushman and Anderson, 1986, 1990; for an exception see Islam et al., in press) or of changes in technological paradigm (Dosi, 1982, 1984; Nelson and Winter, 1982), both of which are of vital concern as documented in a variety of settings (e.g., Christensen and Rosenbloom, 1995; Foster, 1986; Nieto et al., 1998; Schilling and Esmundo, 2009). We contend that the implications of such fundamental changes are also relevant for open innovation in networks, which are both agents of and affected by innovation and change (Sydow et al. 2012). For instance, coordinating open innovation in networks demands specific collaborative mechanisms and practices. This is made more difficult because there is not always a central ‘network orchestrator’ (Dhanaraj and Parkhe, 2006) present, as previously suggested (e.g., Maula et al., 2006; for an overview on leadership in networks see Müller-Seitz, 2102). Instead, in some cases at least, networks are heterarchic in that no one organization formally presides over the network or has any formal quasi-hierarchical control. This is particularly obvious when organizations form R&D consortia, and with their help pursue a given technological path (Roelofsen et al., 2011; Müller-Seitz and Sydow, 2012; Sydow et al., 2012). In these cases governance structures must support collaborative decision-making in order to operate effectively. To exploit the advantages of innovation by such consortia at the level of the ‘whole network’ (Provan et al., 2007), consensus building is critical and must be driven by management (Sydow et al., 2016). Here we discuss the specific question of how organizations work together in a heterarchical network on more or less equal terms in the face of extreme technological uncertainty as might apply when one technological paradigm comes to an end and the 3

industry engages in an intense search for its replacement. Like the studies of open innovation mentioned above, previous research on technological discontinuities has not only adopted a generally firm-centric perspective but has also explored the role of competition (e.g., Nelson and Winter, 1982) – at the expense of cooperative approaches such as collaboration in networks. However, this approach no longer reflects the situation in many industries, e.g., biotechnology (Powell et al., 1996) or the semiconductor manufacturing tool industry (Sydow et al., 2012), where collaboration between organizations via networks is very common and allows the joint assessment of uncertainties and cost issues (Möllering and Müller-Seitz, 2018). This central observation leads us to the following guiding research question: How do heterarchical open innovation networks master transitions across technological discontinuities? We answer this research question by investigating how the heterarchical semiconductor industry network ITRS (International Technology Roadmap for Semiconductors) is facing the transition from one technological paradigm (‘CMOS’) to another as yet unknown technological paradigm (‘Beyond CMOS’). This technological discontinuity is particularly challenging because not only must a new technological paradigm be identified, but relevant partners, as well as the performance and relational risks involved in the new collaboration (Das and Teng, 1999), are also unknown. Moreover, the well-established practice of roadmapping has some limitations when applied to the unknown technological paradigm (Möllering and Müller-Seitz, 2018; Müller-Seitz and Sydow, 2012). ITRS members therefore deem open innovation to be necessary at the network level because no single organization is capable of launching the unknown technological paradigm on its own. We herein contribute to the literature on this topic by introducing the concept of 'stretching practices', defined as an attempt to expand an established practice to an unknown technological and organizational landscape. We further contend that open innovation is 4

currently being practiced collaboratively at the network level in order to address the technological discontinuity. In so doing, we refine previous firm-centric conceptions dealing with transitions between technological paradigms (Dosi, 1982), as well as other related debates relating to S-curves (Geroski, 2000), which focus on rivalry between organizations at the expense of more collaborative efforts to face technological paradigm transitions through the use of networks. Thereby we also lay decided emphasis on the role of uncertainty, which resonates with previous research on technological discontinuities and crises (Archibugi et al., 2013; Roper and Tapinos, 2016; Sydow et al., 2013).

2. Towards an open innovation approach at the network level to address technological discontinuities 2.1 Moving beyond firm-centric open innovation While innovations were traditionally pursued by firms acting alone, recent evidence shows that there is now an increasing reliance on external sources for new ideas. It is against this background that Chesbrough (2003) coined the term ‘open innovation’ to signify collaboration with external partners by managing the in- and outbound flows of knowledge in order to foster intraorganizational innovation and generate opportunities in the market. The findings of this seminal publication have been confirmed by a number of authors in different sectors and settings (e.g., Chesbrough and Crowther, 2006; Chesbrough et al., 2006; Gassmann and Reepmeyer, 2005; Lichtenthaler, 2011), and the central themes have gained increasing attention in managerial practice and academia. While previous research has primarily served our understanding of managing open innovation from the perspective of a single firm interacting with external partners (e.g., Lichtenthaler, 2011; Vanhaverbeke, 2006; West et al., 2006), relatively little attention has

5

been devoted to how the management of open innovation unfolds at the network level (cf. Provan et al., 2007; Nambisan and Sawhney, 2011). Where networks have been analyzed from a perspective of open innovation (Gassmann et al., 2010; Maula et al., 2006; Vanhaverbeke, 2006; Vanhaverbeke and Cloodt, 2006; West, Vanhaverbeke and Chesbrough, 2006), the focus tends to be on hierarchical networks or on how a single organization can benefit from engagement with networks. For instance, Enkel’s (2010) study of a European research network explores the personal and organizational attributes required if a firm is to benefit from engagement with an interorganizational network. Along similar lines, other researchers have sought to explain the processes of intraorganizational change as a result of engaging in networks from the perspective of a single firm (e.g., Dittrich and Duysters, 2007, for the case of Nokia). We further submit that open innovation within networks is relevant for both the network and its members, because networks are both agents of, and affected by, organizational and interorganizational change (Sydow et al. 2012). For instance, managing the boundaries of the network – in addition to organizational boundaries (Santos and Eisenhardt, 2005) – is becoming critical, because the effective organization of the network over time implies opening and closing these boundaries; in other words, deciding on who may be admitted as a member. Analyzing the management of technological innovation at the network level of analysis thus appears timely and relevant for open innovation management practitioners and academics alike (e.g., Gassmann et al., 2010; Vanhaverbeke, 2006; West et al., 2006). The coordination of open innovation activities at the level of whole networks demands specific collaborative mechanisms and practices, an observation that holds particularly true for heterarchical networks where no leading central network orchestrator is present (Dhanaraj and Parkhe, 2006), in contrast to hierarchical networks (e.g., Maula et al., 2006). The need for such practices becomes especially clear when network partners try to pursue different 6

technological paths (Roelofsen et al., 2011; Sydow et al., 2012). In order to exploit the advantages of innovation in such heterarchical networks, consensus is key and must be sought wherever possible (Müller-Seitz, 2012).

2.2 Collaborative approaches to dealing with technological discontinuities Studies of technological discontinuities have highlighted the difficulties of managing the transition from one technological path or paradigm to another (Dosi, 1982; 1984; Roper and Tapinos, 2016; Tushman and Anderson, 1990). As in open innovation research in general, previous research on discontinuities has primarily adopted a firm-centric perspective, assuming a degree of rivalry between the organizations involved. In the evolutionary approach of Nelson and Winter (1982), longitudinal analyses are used to show that over time industries naturally select the technologies and routines that are both (1) the most commonly adopted and (2) best suited to their needs. In common with much of the thinking at the time, these authors assumed that the selective forces of the market work on the single firm, rather than on interorganizational constellations of firms or collaborative networks.

The same

assumptions are made in research on S-curves, which in their basic form describe the performance development of a given technology in relation to the effort invested, but – because of their relatedness to diffusion and life cycle models (Nieto et al., 1998) – can also be used to understand the process of technological change and to analyze the long-term spread of a technology in a market (e.g., Foster, 1986; Geroski, 2000; Schilling and Esmundo, 2009). While this research views the beginning of a new technological paradigm as the start of a new S-curve whose location and slope relative to the established paradigm are not yet known, the question of how the transition from one technological paradigm to another is managed at the level of firms or industries – not to mention networks – is left unanswered.

7

Although these studies have advanced our understanding of the ways of dealing with technological discontinuities, they tend to ignore more collaborative approaches to the problem. We contend that this limits our understanding of how many industries face today's technological discontinuities; such industries include the biotechnology (Powell et al., 1996) or semiconductor industries (Sydow et al., 2012), in which networks are the dominant form of economic governance. These networks are critical insofar as they help to pool complementary resources, reduce the costs of research and development, and manage technological uncertainties. This particular form of governance ought therefore to be incorporated into conceptions of how to face technological discontinuities, in particular from a practice perspective.

2.3 A practice-based perspective on dealing with technological uncertainties We advance a practice-based perspective, in order to analyze how open innovation unfolds at a network level in the face of technological discontinuities (cf. Sydow et al., 2016 for network-related practices and in general Vaara, Whittington, 2012). In line with Giddens (1984), we perceive practices to be ordered, recurring social activities that are relatively stable in time and space, and are shared by different actors. They are not single, isolated occurrences, but are rather part of an ongoing stream of activities in a particular context. The major advantage of such a processual conception is that it allows us to consider activities as they are constrained and enabled by rules and resources, and in turn reproduced or transformed through the use of practices. Furthermore, such an analytical conception allows a focus on practice-immanent dynamics and contradictions rather than on stability and equilibrium (cf. Nicolini, 2013: 44-53). In connection with facing technological discontinuities, it is noteworthy that such a conception assumes that the practices of individuals as well as of collective agents such as 8

organizations or even networks are routinized to some extent; as routines, they are considered to be rooted in the practical rather than the discursive consciousness (Giddens, 1984: 5-14). Only if triggered by an unexpected event or problem, or by the intervention of a third party, is the routine character of a practice likely to be questioned and consequently brought to the discursive realm.

3. Research context and methods The semiconductor industry was selected as the setting of this study because it is characterized both by a dominance of innovation networks as a form of economic governance and by an extremely high degree of technological uncertainty (Browning and Shetler, 2000; Möllering and Müller-Seitz, 2018; Müller-Seitz and Sydow, 2012; Sydow et al., 2012). The latter stems mainly – but by no means exclusively – from the inability to predict the technological paradigm that will succeed the current CMOS-related technological path, i.e., the contested arena typically referred to as ‘Beyond CMOS’. Organizations in the semiconductor industry are still investing in ways to improve the efficiency of current CMOS technologies. These efforts may be thought of as ‘scaling’, referring to Moore’s (1965) law predicting that the number of transistors that can be fitted on an integrated circuit doubles approximately every 18 months. Although this development has been plagued by technological and organizational uncertainties (Sydow et al., 2012; Schubert et al., 2013), it is not at all comparable in this respect to the Beyond CMOS alternatives that are currently envisaged, which question almost every technological aspect, and by extension, the activities of semiconductor industry actors. Moreover, the organizations involved must maintain a constant vigil in order to maintain their current dominant CMOS-related position in the industry, and are fearful of becoming irrelevant or at least marginalized as a result of

9

unforeseen technological discontinuities resulting from Beyond CMOS technologies. Organizations from the CMOS paradigm therefore collaborate with partners to reduce technological uncertainty, although in so doing they may merely end up trading environmental for systemic uncertainty (Sydow et al., 2013). The key strategic mechanism used to deal with these uncertainties in practice is the ITRS, which is both an interorganizational network and a written document displaying the future technological standards of the same name. Industry actors consistently refer to the ITRS document as the most important guideline in the industry (Schaller, 2004), produced for the first time at a national level in 1992 for US organizations, with international members being allowed to participate since 1999. The importance of the ITRS stems from the guidance it offers to actors, given that the organizations concerned (from it and from the industry as a whole) integrate the ITRS into their own strategic planning (ITRS, 2012). Thus ‘being in line’ with the roadmap offers valuable orientation for all actors who then rely on each other in their attempts to reduce development costs and to handle technological uncertainty. ---------------------------------------------INSERT TABLE I ABOUT HERE ---------------------------------------------The ITRS is not only an artifact (the roadmap itself) but also represents an innovation network by the same name that can be described as a whole network (Provan et al., 2007) made up of members (nodes) linked together by virtue of their assignment to the ITRS, in addition to their work for their own organizations. The link between these (currently 140) members from 32 different organizations and one other network (the SEMATECH consortium, see Table I for an overview), is maintained via regular meetings in which actors define future technological milestones more or less in consensus, as well as reproducing the jointly formulated roadmap itself (Schaller, 2004; Sydow et al., 2012). This we define to be 10

the practice of roadmapping. In such meetings, by referring to the ITRS roadmap, each organization participating in the ITRS network reflexively monitors the activities of the other members, thereby constantly redrafting the roadmap document. In Technical Working Groups (TWGs), ITRS members are organized around a specific technological option within a specific technological paradigm.3 The chair of a TWG coordinates the sub-network and decides whether or not a new member is allowed to participate in the TWG and network activities. Eligibility for TWG membership is loosely defined based on required expertise and regional affiliation. The activities of the different TWGs are approved and overseen by the International Roadmap Committee. This committee consists of ‘spokespersons’ from five regional associations,4 each of which assigns two or three representatives to the committee from different organizations. The key activity pursued both in the TWG sub-networks but also in the network as a whole is the identification of future technological milestones in the ITRS (document) in relation not only to CMOS but also to Beyond CMOS, targeting in particular the migration from CMOS to Beyond CMOS. To this end, organizational representatives meet three times a year at global ITRS meetings as well as throughout the year in different regional and technology-

3

There is a formal differentiation between TWGs and Domestic / Regional Technological Working Groups

(DTWGs). While the former are geared towards a technological option, combining expertise from participating organizations on a global scale, the latter are restricted to regional (e.g., related to Europe) or domestic (e.g., US) representatives. This leads to a mirroring of the TWG structure. This differentiation is not discussed further here, because the differentiation adds no value to our argument; furthermore, respondents consistently viewed this differentiation as obsolete in practice, in favour of the overarching TWGs; furthermore in some cases there are no longer enough organizations to represent all the technological areas in the DTWGs (e.g., in Europe). 4

These associations are the European Semiconductor Industry Association (ESIA), the Japan Electronics and

Information Technology Industries Association (JEITA), the Korean Semiconductor Industry Association (KSIA), the Taiwan Semiconductor Industry Association (TSIA), and the United States Semiconductor Industry Association (SIA). 11

related meetings or via IT-enabled exchange forums (e.g., videoconferencing). Officially, the TWGs strive for ‘chair-led consensus’ across the sub-network, although in practice the TWG chair acts as primus inter pares and ‘manufactures’ consent whenever disagreements arise (Buroway, 1982; cf. also Müller-Seitz and Sydow 2012 on how Intel occasionally tried acting as a primus inter pares). For technological themes relevant to more than one TWG, so-called ‘Cross-cut TWGs’ exist, which in effect impose – where applicable – a partial matrixorganizational structure on the sub-networks. In order to analyze how firms work together in networks in the face of high technological uncertainty, we adopt an interpretative research methodology that allows us to capture respondents’ perspectives and practices (Lincoln and Guba, 1985). We chose a longitudinal case-study approach, because this allows us to generate novel insights into how actors align their activities with regard to existing, as well as unknown, technological paradigms (Yin, 2009). Following previous work in optical technologies that sensitized us to the challenges faced by the semiconductor industry, our present study is concerned with two related projects (2004-2010 and 2010-2013) that examined the way complex system technologies are created and extended in the semiconductor industry with the assistance of networks in search of innovative and unknown manufacturing technologies.

3.1 Data collection Our data cover the 21-year period following the inception of the US-based National Technology Roadmap for Semiconductors (NTRS) as an industry-wide roadmap in 1992 and its global successor, the ITRS. Data collection took part in three separate research projects in which both authors were involved. One project dealt with the way interorganizational collaboration is pursued in the quest for the next generation of lithography in the CMOS arena (see Möllering and Müller-Seitz, 2018; Schubert et al., 2013; Sydow et al., 2012). Building on 12

this, the second project dealt with network leadership (conceptually: Müller-Seitz, 2012b; empirically: Müller-Seitz and Sydow, 2012), knowledge management practices at the organizational (Müller-Seitz, Güttel, 2012) and network levels (Müller-Seitz, 2012a). These two projects informed the present study in terms of a structuration understanding of how to lead networks or engage in knowledge management practices in general and roadmapping practices in particular. We built upon these insights to embark on our inquiry in the ways in which coordination is practiced in networks in the face of the transition from CMOS to Beyond CMOS. Apart from our preliminary collection of data from secondary sources, four main sources were used for triangulation purposes to heighten the validity of the constructs, as well as to prevent post-hoc rationalization and potential bias (Yin, 2009), namely field documents, semi-structured interviews, panel-interviews and participant observation. First, we collected a broad range of field documents, including online materials, archival data reproduced via databases, and other documents issued by semiconductor industry actors (e.g., annual reports) for this study, which represent a different data set than for those studies mentioned above. Second, we conducted 143 semi-structured interviews (on average: 60-90 min.) with semiconductor industry experts and senior executives in Europe, the USA, and Japan, the majority of which were conducted during the first project and focused almost exclusively on the CMOS paradigm and possible on-path changes. Among the interviewees were representatives of SEMATECH (54, by far the most important consortium in the field), suppliers (35), chip manufacturers (17), and other networks and consortia (8), as well as senior civil servants (6), representatives of research laboratories (10), and consultants (1), all of which groups assign members to the ITRS. Among these interviewees were, on two occasions, the head of the NTRS/ITRS and the relevant chairpersons of the different TWGs and regional associations. For the present study, it was most useful that in April 2011 two 13

members of the research team were invited by a member of the ITRS board to conduct 15 interviews with respondents from different organizations in the network. This round of interviews was particularly geared towards identifying strategies of how to face technological uncertainty ensuing from the expected transition from the CMOS technological paradigm to an as yet unknown technological paradigm. In total, the least 30 interviews by and large focused only on the topics presented in this manuscript and were conducted by one of the coauthors alone. Third, in the CMOS-focused first part of the study, five semiconductor industry experts were interviewed in the form of a panel between 2007 and 2010. Despite the focus of the panel interviews on the development of CMOS, there were also some questions related to Beyond CMOS, which raised our interest and led to the aforementioned set of interviews with a special interest into the present topic of open innovation stretching practices in the face of uncertainty. Fourth, we draw on material obtained from the observations of participants at workshops and conferences, in the latter case both from participation at conferences by one of the authors between 2009-2011, and from an analysis of online available conference presentations, slides and public announcements gathered during both project phases during these meetings. The observations of participants allowed us to conduct impromptu interviews lasting 5-60 minutes, but these were not transcribed. Finally, we engaged in follow-up interviews and e-mail correspondence with key respondents, and in discussions with five US and three European scholars from the fields of strategic management and organizational sociology, in the interests of validation. In this context, we particularly benefitted from a 110-minute discussion on the content of our study with three ITRS board members. Together with our prolonged period of field data collection, this reduced the risk of fundamental misinterpretation of the data. 14

3.2 Data analysis Although our data analysis did not occur chronologically, it can be divided roughly into three stages. In the first stage we collected all data in a case study database in order to strengthen its reliability (Yin, 2009). Our whole analysis was based on the ‘raw data’ obtained from 160 pages of field notes, about 1,600 pages of interview transcripts, and roughly triple that amount of archival material and ITRS conference data. In stage two, we drafted condensed descriptions of the ways in which the different network members interacted, particularly at the regular meetings of the ITRS network and in the ongoing revision of the ITRS document itself. We also tracked the development of technological options within and between different technological paradigms. In stage three we condensed our empirical data, as is common practice in longitudinal and in-depth case-study research (Yin, 2009). To this end we entered all our ‘raw data’ into atlas.ti, a software tool for analyzing qualitative data. We started coding our data by seeking ways in which the ITRS network coordinated its activities with regard to the CMOS, as well as the Beyond CMOS paradigms; in other words we attempted to identify ways of dealing with both technological paradigms, which resulted in our emerging conception of 'stretching'. At first we remained open to the multifaceted options available for pursuing these activities, and we chose an inductive approach as a means of labeling first-order categories concerning the different collaborative activities. We then attempted to construct mutually exclusive second-order themes and grouped these hierarchically. This procedure allowed the collapse of first-order categories into second-order themes that represented more abstract and researcherinduced interpretations. Thereafter, the second-order themes were subsumed under the two third-order themes, which constituted the core of our conception of stretching practices in the face of technological discontinuities. 15

4. Collaborative network coordination via roadmapping Our analysis suggests that roadmapping – understood as constantly producing and reproducing jointly defined future technological milestones – is the key activity of the ITRS in the semiconductor industry in its management of technological discontinuities in general and in facing technological uncertainty in particular. We show that roadmapping is used in connection with CMOS, and given the experience thus gained, is also applied in a similar fashion to the unknown future technological paradigm termed Beyond CMOS. ----------------------------------------------INSERT FIGURE 1 ABOUT HERE -----------------------------------------------Our analysis of the ITRS data allowed us to sketch the development of multiple technological options and their accompanying TWGs over time (cf. Figure 1). This served as basis for understanding the way roadmapping has already been used to deal with the technological uncertainties associated with both the CMOS and the Beyond CMOS technological paradigms.

4.1 Roadmapping in the CMOS paradigm From the start, the NTRS and ITRS were geared towards CMOS-related scaling, and offering guidance for R&D investment decisions for the industry as a whole. Table II offers illustrative evidence of the value of roadmapping in relation to the CMOS paradigm and for Beyond CMOS (cf. section 4.2). Generally, this guidance consists of setting technological milestones by consensus, transforming non-calculable technological uncertainty into calculable

16

technological risk from actors’ perspectives (I5-118; I-136). According to a former Intel CEO (I-113), the guidance is a key element of roadmapping, as actors must remain in line with Moore’s law in order to maintain their innovative image for investors, the media and the public. Thus the ITRS network targets “present[ing an] industry-wide consensus on the ’best current estimate’ of the industry’s research and development needs” (ITRS, 2012: 1). The success of roadmapping by the ITRS can therefore be attributed to the consensus-oriented nature of the document, as consistently confirmed by respondents. An ITRS board member called it a “neutral book” (I-125). ----------------------------------------------INSERT TABLE II ABOUT HERE -----------------------------------------------Additionally, the ITRS makes it easier to handle the technological uncertainties involved in following the comparatively well understood CMOS path (I-133). To this end, basic research programmes are launched by the ITRS in connection with different partners. For instance, a respondent stated that companies use “industry and government support to make the infrastructure realistic” (I-58). At an ITRS meeting a TWG chair made an open call for participation in an upcoming project that targeted the reduction of technological risks concerning a specific CMOS technology. Although the CMOS paradigm is rather better understood than the Beyond CMOS paradigm, several technological uncertainties remain (Sydow et al., 2012; Lange et al., 2013). Take, for example, the quest for a next generation lithography where a dominant technological path is sought within the CMOS paradigm. Currently, the most likely candidate is EUV. However, it is by no means certain if or when this path will ever succeed. Another

5

Here and elsewhere “I-” refers to the interview number, for instance, I-118 indicates interview 118. 17

example is the timing of the end of the CMOS paradigm. Novel technological paths are often initiated to complement existing efforts to follow Moore’s law. One of these is the so-called ‘More than Moore’ path, which “addresses an emerging category of devices that incorporate functionalities that do not necessarily scale according to Moore's Law [as well as] new devices that will realize some ‘Beyond CMOS’ capabilities” (ITRS, 2012: 2). Illustrative data highlighting the role of technological uncertainty in the case of roadmapping for the CMOS paradigm are shown in Table III. ---------------------------------------------INSERT TABLE III ABOUT HERE ---------------------------------------------While some TWGs have emerged over time within the CMOS network, others have ceased to exist, depending on the relevance and viability of the technological options considered. For instance, the TWG on Defect Reduction (geared towards reducing errors in the manufacture of semiconductors) stopped meeting in 2001, although most of its members redefined their scope of activities and reinstalled this particular TWG later on as the Yield Enhancement TWG, which broadened in scope to include the detection and sources of defects, in addition to their reduction. The changes were legitimized at the field-wide, Fifth NGL Sematech Workshop in 2001, at which actors had to find some consensus on the technological options they considered most successful within the CMOS paradigm. The reason for this forcing of the consensus was that the field as such could not afford to pursue multiple CMOS options, so they aimed to transform genuine uncertainty into more or less calculable risks by defining (allegedly) measurable objectives to reduce technological options. In fact, EUV turned out to be the consensus and was therefore described as the desperately needed “game-changing technology” (I-103) sought by the entire field.

18

The foregoing characteristic is typical of the ITRS as it tries to strive for consensus in the process of roadmapping. Building consensus is often difficult because “personal interests, company interests, the economic climate and competitive dynamics” (I-128) influence roadmapping across organizations. Hence it is not surprising that TWG chairs tend to step in if no consensus arises and “gently press” (I-98), as one respondent put it, for consensus in the general direction (s)he favors. In one case of disagreement, the same respondent reported that the respective TWG chair asked TWG members whether they “would please at least sign the document as a participant”, as if to record a minimal degree of consensus. The development of technologies within the CMOS paradigm such as EUV is challenged by uncertainty, however. For instance, despite the status of EUV as the most favored nextgeneration lithography, its introduction is repeatedly questioned due to the critical technical issues that must be resolved (I-84; I-129; I-136). If no resolution is possible, this could result in organizations falling behind Moore’s law, which has severe financial implications; shareholders and industry observers tend to punish actors who lack the capacity to innovate (I-98). As a result, it is unsurprising that actors exchange knowledge in this pre-competitive global arena by roadmapping, in order to align the infrastructure necessary to stay in line with Moore’s law, and substantial financial resources are devoted to this end.

4.2 Challenges for roadmapping Beyond CMOS Roadmapping in the ITRS is not bound solely to engagement with organizations working on CMOS-related technologies. Organizational actors also recognize the need to open up the research agenda to Beyond CMOS and to collaborate with partners not yet engaged in the development of the new technologies. Roadmapping could therefore lead to an opening up of the collaborative network – as well as of the innovation process – by the restructuring of the ITRS as a network organization for roadmapping. Table IV offers illustrative evidence of the 19

different forms of uncertainty in the process, related to (1) the as yet unidentified novel technological paradigm, as well as (2) the organizations to engage with, and (3) questioning roadmapping as a successful procedure per se. -----------------------------------------------INSERT TABLE IV ABOUT HERE -----------------------------------------------To face these three forms of uncertainty, the ITRS tries to apply roadmapping as pursued for CMOS to the Beyond CMOS arena (cf. Table II for the way roadmapping is extended to Beyond CMOS). There are two reasons for this: first, roadmapping proved to be very successful for CMOS. During an ITRS meeting we interviewed a member of the steering committee who confirmed that “there are several paths […] being explored. So it is of interest for all the players in the industry to try to replicate the success we had with Moore's law [i.e. CMOS scaling]“ (I-129). Second, there seems to be no alternative, more promising approach than roadmapping. To that end, an interviewee at an ITRS meeting mentioned that “the industry keeps pumping a lot of money into trying to find […] a Beyond CMOS option”. During another ITRS meeting a respondent stated that “we [as ITRS network] can keep the style of the ITRS, but the contents of the ITRS is changing, depending on which technology is promising or not”. In a similar vein, one board member highlighted the evolving nature of the ITRS, when he mentioned that the ITRS “keeps growing and it [the ITRS as a document] is simply like the diary of the industry. You know, it is repository of everything that is happening within the industry” (I-125). We now discuss the three forms of uncertainty in more detail. (1) Technological uncertainty is present insofar as the ITRS has no distilled set of technological options as it has for CMOS technologies (I-124). At best, there are “many candidates for Beyond CMOS devices [which is] probably the most pressing thing now” (I20

126). As a result, the ITRS makes an attempt to “look at the big trends” (I-129) that might be relevant in the future. Given the extreme uncertainty surrounding the future technological landscape of Beyond CMOS, an interviewee mentioned that the ITRS and its members need to identify “opportunities outside [their] comfort zone” (I-121). To this end, exchanges with adjacent communities are initiated as a key measure. Previously untapped technological areas and collaborators could thus start to become relevant, e.g., nanoelectronics or biotechnology (I-128). The aim is thus to Determine which, if any, current approaches to providing a “Beyond CMOS” informationprocessing technolog[ies are] ready for more detailed roadmapping and enhanced investment (Hutchby, 2008: 5). At the level of the ITRS network, another key measure for addressing technological uncertainty was the introduction of two TWGs (in 2002 and 2005) to the ITRS network, devoted entirely to challenges related to Beyond CMOS, entitled Emerging Research Devices and Emerging Research Materials. Both chapters are considered “incubators for future potential technology solutions” (Herr, 2011: 26). This was the result of a collective, networkwide agreement on the need to invest in these new areas as a means of facing the uncertainty regarding alternative technological paths following the end of the CMOS paradigm. The ITRS and its members were afraid that “we will have nothing to guide our research” (I-129), which is why roadmapping was applied to the Beyond CMOS arena. Technological uncertainty is also shown very clearly during the course of roadmapping, in that the roadmap document shows the most uncertain areas in red, whereas other technological options and their accompanying technological parameters – in particular within the CMOS paradigm – are highlighted in white (for technological options that ‘only’ need to be optimized), or yellow (for technological options that are known, but need further development). As for the highly uncertain technological solutions, they represent as defined

21

by the ITRS (2007) “Red brick walls [designating] manufacturable solutions [that] are NOT known”. In this way the ITRS tries to continue roadmapping as for the CMOS paradigm while incorporating – by means of different colors within the ITRS document – the potentially relevant technological paths of the Beyond CMOS paradigm. (2) Partner-related uncertainty ensues from technological uncertainty, as roadmapping is limited by an inability to foresee who the collaborators are. Actors in the existing semiconductor industry are convinced of their inability to face the challenges of Beyond CMOS on their own, hence the need to open up to new organizations from other industries (I121). One board member stated during an ITRS meeting “if you ask about the players there, this can be almost anyone”. Opening up the activities of the network causes is aggravation, because it also implies that “the number of players is going to decrease, there may be only few left in a few years from now” (I-122). As a result, due to the unclear technological as well as organizational, i.e., actor- or network-related landscapes, activities geared to Beyond CMOS appear somewhat experimental in nature (I-127). One of our respondents highlighted the need to open up the network organization, as well as the difficulties inherent in so doing, when he pointed out: I would like to see much more dynamism in many ways […] It is still the old club [i.e. existing and established semiconductor industry actors] which I perceive to be behind the times. Everyone has found his role […] everyone knows how everything works [concerning CMOS-related technologies] and they focus upon that […] it would really need some ‘fresh blood’ (I-118). Nonetheless, and as shown in Figure 1, even the CMOS-related actors acknowledge the increasing relevance of Beyond CMOS, given the presumption that CMOS will come to an end and a future technological paradigm must be envisioned in parallel. It is rather presumed that the CMOS and Beyond CMOS technological options might be interconnected within the future technological path (I-121; I-129). Some TWGs have therefore also convened for so22

called Technology Pacing-sessions, trying to establish links across CMOS-related TWGs as well as engaging with Beyond CMOS-related TWGs, thereby developing the network further. (3) Finally, we also discovered procedural uncertainty surrounding the roadmapping per se, in questioning the way actors collaborate in the face of Beyond CMOS and how they will develop the technology. The consensus-driven nature of the roadmapping was deemed a key asset for the CMOS paradigm, but given the lack of a clear technological path, roadmapping as a ‘tool’ or procedure may be of more limited use (I-134). This point is perhaps best illustrated by the continuous pursuit of quantifying technological challenges by assigning measurable variables to them (e.g., to determine defect rates or the accuracy of assessing the maturity of a technology). The tendency to quantify technological characteristics has increased over the years, and respondents suggested that this limits the value of the ITRS with regard to Beyond CMOS, offering little or no support to decision-makers when no quantifiable data exist (I-120). Another concern relates to the question of whether one single roadmap is still adequate. While roadmapping for CMOS at least implicitly assumes that miniaturization in line with Moore’s law leads to a standardized output that customers refine in line with their own purposes, the same assumption does not apply to Beyond CMOS (I-129). Instead, it is assumed that several roadmaps will probably be necessary to accommodate different customer segments, e.g., health or automotive industry organizations (I-134), which is why the organizations are faced with the question of “how is it still possible to develop roadmaps – with an ’s’ – because there will be many of them?” (I-129). In light of the previous observation, roadmapping as it is currently pursued indeed seems to be somewhat limited in value, due also to the extensive and complex nature of the ITRS documentation. At the time of its introduction and in the early years of the ITRS, reading and comprehending the ITRS document was possible for industry experts; now, however, the 23

document has become too long (i.e., > 1,000 pages) and the detail is incomprehensible to most individuals (I-121). This is noteworthy, since it is therefore losing its function as an overview for the industry as a whole. In relation to this, a Beyond CMOS TWG chairman mentioned that “what is […] missing today is more the expertise, the systems expertise … so we have people who are expert in certain domains” (I-129). This problem is aggravated in view of the difficulty for semiconductor industry experts of communicating the ITRS content to other industry representatives with whom they might wish to coordinate their activities with regard to Beyond CMOS (I-129). Indeed it seems impossible to communicate with adjacent or as yet unacquainted scientific communities even with the help of the ITRS. However, there is presently no alternative option for joining forces in the face of technological uncertainty arising from Beyond CMOS. Roadmapping is therefore still deemed to be the most suitable practice for to the ITRS network (I-122; I-134).

5. Stretching practices to enable open innovation at the network level in the face of technological discontinuities 5.1 Towards a concept of stretching practice To cope with the high technology-, partner- and procedure-related uncertainty of an unknown future technological paradigm, ITRS members must expand the existing CMOS roadmap practices for use in planning for the Beyond CMOS arena. Extending the existing practice of roadmapping illustrates what we comprehend as ‘stretching practice’. Generalizing from this observation, we define stretching practice to comprise recurrent activities that aim to transfer a common way of dealing with the risks and uncertainties of an existing (in our case: technological) paradigm to an as yet unknown (technological) paradigm. This definition

24

therefore comprises collaborative efforts across organizational boundaries (as in the case of the ITRS), but may cover intraorganizational contexts as well. In light of the findings discussed above, we suggest that both individual organizations and collectives of organizations rely on rules and resources when stretching practice, in the form of structures established and reproduced for the existing (technological) paradigm (cf. Figure 2). This stretching of practice, if successful, results in transferring a practice to less well known context, allowing organizational actors to act under conditions characterized by fundamental uncertainty, by virtue of their ability to draw upon established stretching practice. ----------------------------------------------INSERT FIGURE 2 ABOUT HERE -----------------------------------------------Because there are no obvious alternative practices for roadmapping, we comprehend the constant production and reproduction of future technological milestones (i.e., roadmapping) to be a practice that is comparatively stable in time and space, and enabled and constrained by structures (Müller-Seitz, 2012a; Müller-Seitz and Sydow, 2012), i.e., rules of signification and legitimization as well as resources of domination (Giddens, 1984). Furthermore, the successful reproduction of this practice depends on a set of sub-practices, e.g., conducting global ITRS meetings annually together with regional TWG meetings in Europe, Asia and the USA; sub-practices that constitute roadmapping by means of updating the ITRS document. In turn, this artifact helps to reproduce the sub-practices. Taken together, this reproduction of the ITRS culminates in developing the network and its relationships across the network members and, to some extent, even developing the organizational field of semiconductor manufacturing. This is achieved by reconvening the same TWG members as for CMOS,

25

while at the same time opening the network up to new partners in new fields related to Beyond CMOS. In the case considered here, this practice of roadmapping thus acquires a ‘stretched’ character, insofar as the present ITRS network is trying to extend its existing means of dealing with technological (and some organizational) uncertainties of the CMOS paradigm to the Beyond CMOS arena. The stretching of these practices is not only enabled and constrained by the rules and resources mentioned above, but can itself also be conceptualized in terms of signification, legitimation and domination (Giddens, 1984). On the one hand, stretching implies sensemaking activities (Weick, 1995) with regard to rules of signification and legitimization that are compatible with those that characterize the network and the field, i.e., they allow for communication and sanctioning in ways that are understood and accepted. On the other hand, stretching practice, which becomes a resource that allows actors to “make a difference to a pre-existing state of affairs or course of events” (Giddens, 1984: 14), relies on the ability of organizational agents to mobilize network or even field resources, in terms of time and money, for instance. To this end, the semi-permeability of the network is reproduced with the help of such stretching practice as network boundaries – for practicing more open innovation – become more porous; that is, while trying to explore the unchartered technological paradigm of Beyond CMOS, new members are included when introducing new TWGs such as Emerging Research Materials and Emerging Research Devices. The term ‘stretching’ has been employed elsewhere with different connotations. Perhaps closest to our conception is that of ‘strategy as stretch’ (Hamel and Prahalad, 1993), although these authors refer to the differences between resources and goals rather than, as in our case, between differing and as yet partially undefinable goals. O’Mahony and Bechky (2006) also make use of the term when they refer to individual career aspirations in terms of bridging existing competencies in order to extend them to new areas. Of particularly note is the work 26

of Sitkin et al. (2011), who focus on ‘stretch goals’, understood not necessarily as two or more differing objectives, but as future conditions that are difficult to obtain. Our own notion of stretching practices is quite similar, but differs firstly because our key interest lies in two competing objectives (CMOS and Beyond CMOS) where the objectives, at least for Beyond CMOS, are unknown and evolving rather than known and clearly specifiable. Secondly, the concept of stretching is applied herein to the interorganizational realm, specifically at the level of the whole network. However, coping with fundamental uncertainty by stretching an existing collaborative practice is accompanied by the risk of ‘overstretching’; that is, the need to invest financial and other resources might become too great to continue pursuing both one existing and one unknown technological paradigm. This risk is not considered in previous conceptions of stretching (e.g., O’Mahony and Bechky, 2006), which seem, at least implicitly, to assume unlimited resources.

5.2 Contributions of this study With our study and the concept of stretching practice, we contribute to the literature on open innovation in the face of technological discontinuities as follows: first, we concentrate on open innovation at the (whole) network level of analysis (Provan et al., 2007) rather than on how a single organization can benefit from engaging in open innovation-related activities with external partners in a network, as most studies of open innovation have done to date (Gassmann et al., 2010; Enkel, 2010; Maula et al., 2006; Vanhaverbeke, 2006; Vanhaverbeke and Cloodt, 2006; West et al., 2006). This is important, because managing whole networks, in particular heterarchical ones, holds particular challenges. Because the ITRS network constitutes a fairly heterarchical setting, our case sensitizes us to the need to engage in collaborative practices across the network to allow “manufactured consent” (Buroway, 1982; 27

see along similar lines the conceptual inquiry by Müller-Seitz, 2012b) as a prerequisite to genuine open innovation at the network level. This is particularly challenging for the ITRS, in which numerous powerful organizations participate but no leading firm exists (in contrast to the SEMATECH consortium for example; cf. Müller-Seitz and Sydow, 2012). Second, in relation to previous research on technological discontinuities (Tushman and Anderson, 1986, 1990) or changes in technological paradigm (Dosi, 1982, 1984; Nelson and Winter, 1982), we refine the insights of these research streams by arguing that such changes are not only relevant for a single organization or an industry, but represent a vital concern for whole networks as well (see for a similar argument Islam et al., in press). This implies a shift from a predominantly firm-centric and ‘winner-takes-all’ perspective to a more consensusoriented view that nevertheless accounts for the political character of networks (Benson, 1975). As discussed, establishing collaborative mechanisms in face of this character is of key importance for heterarchical networks over which no one firm presides. Even if there were to be a central network orchestrator (Dhanaraj and Parkhe, 2006; Nambisan and Sawhney, 2011), and even if network members were in competition, firms would nevertheless opt for a joint, collaborative approach for managing the high technological and organizational uncertainties involved in the transition from one technological paradigm to another unknown one. One additional and very important reason for a more collaborative approach is the strategy of collective financing that complements the individual, firm-centered funding that takes place in this industry, as documented by Lange et al. (2013). Third, our findings and the concept of stretching also refine studies of S-curves (e.g., Foster, 1986; Geroski, 2000; Schilling and Esmundo, 2009) in that we concentrate not only on the technological options at hand, but also on the underlying practice pursued to manage the respective technological and organizational discontinuities. To this end, from our findings we submit that the ‘stretching’ of existing practices might not by itself lead to the new and 28

unknown technological paradigm; rather the transformation follows from the network activities that assist the stretching practice of roadmapping.

6. Concluding remarks Our main aim in this research was to explore how open innovation can be mastered at the network level of analysis in the face of an industry’s potentially dramatic transition from an existing technological paradigm to a new and unknown one. The empirical setting chosen for this study was the global semiconductor industry, which is characterized not only by extremely high levels of uncertainty in general but by alternative technological and organizational options with contradictory, possibly even paradoxical demands. The results of our study suggest that stretching the practice of collaborative roadmapping allows, at least to some extent, the exploitation of existing practices on the present path as well as their transfer to an as yet unknown technological paradigm. Further data on stretching practices at an organizational or even an industry level would help to complement our understanding of the way networks and their members face uncertainty. Such a multi-level effort would nicely complement the focus of the present study at the network level of analysis and could thus influence research in the field of open innovation that still lacks a multi-level component (Maula et al., 2006; Vanhaverbeke et al., 2006; West et al., 2006). Another future avenue of research might involve the capture of the practices of organizational actors coping with extreme uncertainties of different kinds, e.g., when they are intentionally inducing uncertainty (rather than reducing uncertainty as primarily observed in our study) to stimulate radical new solutions to problems. In sum, we submit that further explorations of how networks face technological uncertainty and the practices they employ are needed to comprehend this timely and important open innovation management topic more adequately. 29

References Archibugi, D., Filippetti, A., Frenz, M. 2013. The impact of the economic crisis on innovation: Evidence from Europe. Technological Forecasting and Social Change 80 (7), 1247-1260. Benson, K.J. 1975. The interorganizational network as a political economy. Administrative Science Quarterly, 20, 229-249. Browning, L. D., Beyer, J. M., Shetler, J. C. 1995. Building Cooperation in a Competitive Industry: Sematech and the Semiconductor Industry. Academy of Management Journal, 38(1), 113-151. Browning, L. D., Shetler, J. C. 2000. Sematech: Saving the U.S. Semiconductor Industry. College Station, TX: University Press. Buroway, M. 1982. Manufacturing consent: Changes in the labor process under monopoly capitalism. Chicago: University of Chicago Press. Chesbrough, H. 2003. Open innovation. The new imperative for creating and profiting from technology. Boston: Harvard Business School Press. Chesbrough, H., Crowther, A. K. 2006. Beyond high tech: early adopters of open innovation in other industries. R&D Management 36, 229-236. Chesbrough, H. W., Rosenbloom, R. S. 2002. The role of the business model in capturing value from innovation: evidence from Xerox Corporation's technology spin-off companies. Industrial and Corporate Change 11, 529-555. Christensen, C.M., Rosenbloom, R. S. 1995. Explaining the attacker's advantage: Technological paradigms, organizational dynamics, and the value network. Research Policy 24, 233-257. 30

Das, T. K., Teng, B. 1999. Managing risks in strategic alliances. Academy of Management Executive, 13, 50-62. Dhanaraj, C., Parkhe, A. 2006. Orchestrating innovation networks. Academy of Management Review 31, 659-669. Dittrich, K., Duysters, G. 2007. Networking as a means to strategy change. The case of open innovation in mobile telephony. Journal of Product Innovation Management 24, 510-521. Dittrich, K., Duysters, G., de Man, A. 2007. Strategic repositioning by means of alliance networks: The case of IBM. Research Policy 36, 1496-1511. Dosi, G. 1982. Technological paradigms and technological trajectories: A suggested interpretation of the determinants and directions of technical change. Research Policy 11, 147-162. Dosi, G. 1984. Technical change and industrial transformation. New York: St, Martin’s Press. Enkel, E. 2010. Attributes required for profiting from open innovation in networks. International Journal of Technology Management 52, 344-371. Foster, R. 1986. Innovation: The attacker's advantage. New York: Summit Books. Freeman, C. 1991. Networks of innovators: A synthesis of research issues. Research Policy 20, 499-514. Gassmann, O., Enkel, E., Chesbrough, H. 2010. The future of open innovation. R&D Management 40, 213-221. Gassmann, O., Reepmeyer, G. 2005. Organizing pharmaceutical innovation. From sciencebased knowledge creators to drug-oriented knowledge brokers. Creativity and Innovation Management 14, 233-245. Geroski, P.A. 2000. Models of technology diffusion. Research Policy 29, 603-625. Giddens, A. 1984. The constitution of society. Cambridge: Polity Press. 31

Hamel, G., Prahalad, C. K. 1993. Strategy as stretch and leverage. Harvard Business Review 71, 75-84. Herr, D. J. C. 2011. New technologies & devices structures. Future Fab International 36, 26. Hutchby,

J.

A.

2008.

Emerging

research

devices.

http://www.itrs.net/Links/

2008Summer/Public%20Presentations/PDF/ERD.PDF (accessed on 6 January 2011). Islam, N., Gyoshev, S., Amona, D. In press. External complexities in discontinuous innovation-based R&D projects: Analysis of inter-firm collaborative partnerships that lead to abundance. Technological Forecasting and Social Change. ITRS 2012. International Technology Roadmap for Semiconductors. 2012 Update. Knight, F. H. 1921. Risk, uncertainty and profit. Chicago: University of Chicago Press. Lange, K., Müller-Seitz, G., Sydow, J., Windeler, A. 2013. Financing innovations in uncertain networks—Filling in roadmap gaps in the semiconductor industry. Research Policy, 42, 647-661. Lichtenthaler, U. 2011. Open innovation: Past research, current debates, and future directions. Academy of Management Perspectives 25, 75-93. Maula, M. V., Keil, T., Salmenkaita, J. 2006. Open Innovation in systemic innovation contexts. In H. W. Chesbrough, W. Vanhaverbeke, J. West (Eds.). Open innovation. Researching a new paradigm: 241-257. Oxford: Oxford University Press. Möllering, G., Müller-Seitz, G. (2018). Direction, not destination: Institutional work practices in the face of field-level uncertainty. European Management Journal, 36(1): 28-37. Moore, G. E. 1965. Cramming more components onto integrated circuits. Electronics 38, 11417. Müller-Seitz, G. 2012a. Absorptive and desorptive capacity-related practices at the network level – the case of SEMATECH. R&D Management 42(1), 90-99. 32

Müller-Seitz, G. 2012b. Leadership in interorganizational Networks: A literature review and suggestions for future research. International Journal of Management Reviews 14: 428443. Müller-Seitz, G., Güttel, W. 2014. Towards a choreography of congregating – A practicebased perspective on organizational absorptive capacity in a semiconductor industry consortium. Management Learning 45(4), 477-497. Müller-Seitz, G., Sydow, J. 2012. Maneuvering between networks to lead – a longitudinal case study in the semiconductor industry. Long Range Planning, 45, 105-135. Nambisan, S., Sawhney, M. 2011. Orchestration processes in network-centric innovation: Evidence from the field. Academy of Management Perspectives, 25 (3), 40-58. Nelson, R.R., Winter, S. G. 1982. An evolutionary theory of economic change. Cambridge, Massachusetts: The Belknap Press of Harvard University Press. Necoechea-Mondragón, H., Pineda-Domínguez, D., Pérez-Reveles, L., Soto-Flores, R. 2017. Critical factors for participation in global innovation networks. Empirical evidence from the Mexican nanotechnology sector. Technological Forecasting and Social Change 114, 293-312. Nicolini, D. 2013. Practice theory, work & organizations. Oxford: Oxford University Press. O'Mahony, S., Bechky, B. A. 2006. Stretchwork: Managing the career progression paradox in external labor markets. Academy of Management Journal 49, 918-41. Nieto, M., Lopéz, F., Cruz, F. 1998. Performance analysis of technology using the S curve model: the case of digital signal processing (DSP) technologies. Technovation 18 (6/7), 439-57.

33

Powell, W. W., Koput, K. W., Smith-Doerr, L. 1996. Interorganizational collaboration and the locus of innovation: Networks of learning in biotechnology. Administrative Science Quarterly 41, 116-45. Provan, K. G., Fish, A., Sydow, J. 2007. Interorganizational networks at the network level: A review of the empirical literature on whole networks. Journal of Management 33, 479-516. Roelofsen, A., Boon, W. P. C., Kloet, R. R., Broerse, J. E. W. 2011. Stakeholder interaction within research consortia on emerging technologies: Learning how and what? Research Policy 40, 341-354. Roper, S., Tapinos, E. 2016. Taking risks in the face of uncertainty: An exploratory analysis of green innovation. Technological Forecasting and Social Change 112, 357-363. Santos, F. M., Eisenhardt, K. M. 2005. Organizational boundaries and theories of organization. Organization Science 16, 491-508. Schaller, R. R. 2004. Technological innovation in the semiconductor industry: A case study of the

international

technology

roadmap

for

semiconductors.

http://www.xecu.net/schaller/schaller_dissertation_2004.pdf (accessed on 10 June 2008). Schilling, M. A., Esmundo, M. 2009. Technology S-curves in renewable energy alternatives: Analysis and implications for industry and government. Energy Policy 37, 1767-1781. Schubert, C., Sydow, J., Windeler, A. 2013. The means of managing momentum. Bridging technological paths and organisational fields. Research Policy 42, 1389-1405. Sitkin, S. B., See, K. E., Miller, C. C., Lawless, M. W., Carton, A. M. 2011. The paradox of stretch goals: Organizations in pursuit of the seemingly impossible. Academy of Management Review 36, 544-66. Strauss, A., Corbin, J. 1998. Basics of qualitative research. Techniques and procedures for developing grounded theory, Thousand Oaks: Sage. 34

Sydow, J., Müller-Seitz, G., Provan, K. 2013. Managing uncertainty in alliances and networks: From governance to practice. In T.K. Das (Ed.): Managing knowledge in strategic alliances. Information Age Publishing: Charlotte/North Carolina, 1-43. Sydow, J., Schüßler, E., Müller-Seitz, G. 2016. Managing Interorganizational Relations – Debates and Cases, Palgrave / Macmillan Publishers. Sydow, J., Windeler, A., Schubert, C. Möllering, G. 2012. Organizing R&D consortia for path creation and extension: The case of semiconductor manufacturing technologies. Organization Studies 33, 907-936. Tushman, M. L., Anderson, P. 1986. Technological discontinuities and organizational environments. Administrative Science Quarterly 31, 439-465. Tushman, M. L., Anderson, P. 1990. Technological discontinuities and dominant designs: A cyclical model of technological change. Administrative Science Quarterly 35, 604-633. Vaara, E., Whittington, R. 2012. Strategy as Practice. Taking Social Practices Seriously. Academy of Management Annals 6, 285-336. van der Valk, T., Chappin, M.M.H., Gijsbers, G.W. 2011. Evaluating innovation networks in emerging technologies. Technological Forecasting and Social Change 78(1), 25-39. Vanhaverbeke, W. 2006. The interorganizational context of open innovation. In H. W. Chesbrough, W. Vanhaverbeke, J. West (Ed.), Open innovation. Researching a new paradigm: 205-219. Oxford: Oxford University Press. Vanhaverbeke, W., Cloodt, M. 2006. Open innovation in value networks. In H. W. Chesbrough, W. Vanhaverbeke, J. West (Ed.), Open innovation. Researching a new paradigm: 258-281. Oxford: Oxford University Press. von Hippel, E. 1986. Lead users: A source of novel product concepts. Management Science 32, 791-805. 35

Weick, K. E. 1995. Sensemaking in organizations. Thousand Oaks: Sage. West, J., Vanhaverbeke, W., Chesbrough, H. 2006. Open innovation: a research agenda. In H. W. Chesbrough, W. Vanhaverbeke, J. West (Ed.), Open innovation. Researching a new paradigm: 285-307. Oxford: Oxford University Press. Yin, R. K. 2009. Case Study Research. 4th ed., Thousand Oaks: Sage.

36

APPENDIX TABLE I: Persons assigned to the ITRS network by members, 2011-12. Participating member Organization

Name

IRC representative

3MTS Applied Materials ASEUS Axcelis Cabot CNSE Consultantcy

Cymer Dupont Fraunhofer Freescale FSI GlobalFoundries Hynix

Korea

IBM

Intel

Micron Microsystems MXIC NIST

Panasonic Renesas Samsung

USA

Taiwan

Korea

Sharp Solid State Solutions Semiconductor Research Corporation

STMicroelectronics Texas Instruments

Europe USA

Texas State Toshiba

Japan

UCSD UMC Network

COSAR SEMATECH

Sum

34

Korea USA

TWG - Number of persons participating

TWG - Name Assembly & Packaging Yield Enhancement Assembly & Packaging Front End Processes Interconnect Metrology Test & Test Equipment Emerging Research Materials Assembly & Packaging Lithography Environment, Safety & Health Yield Enhancement Modeling & Simulation Wireless Front End Processes Lithography Assembly & Packaging Emerging Research Materials Factory Integration Front End Processes Interconnect Lithography Metrology Modeling & Simulation Process and Integrated Device Structures System Drivers and Design Test Yield Enhancement System Drivers and Design Wireless Process and Integrated Device Structures Factory Integration Environment, Safety & Health Emerging Research Materials Test & Test Equipment System Drivers and Design Process and Integrated Device Structures Wireless Metrology Emerging Research Materials / Emerging Research Devices Metrology Process and Integrated Device Structures Process and Integrated Device Structures KSIA staff Assembly & Packaging Design Emerging Research Materials / Emerging Research Devices Factory Integration Front End Processes Interconnect Lithography Metrology Modeling & Simulation Process and Integrated Device Structures Test Yield Enhancement Interconnect Interconnect Process and Integrated Device Structures Emerging Research Devices Emerging Research Materials Process and Integrated Device Structures Process and Integrated Device Structures Micro-Electro Mechanical Systems Emerging Research Devices Front End Processes Process and Integrated Device Structures Process and Integrated Device Structures System Drivers and Design Design Process and Integrated Device Structures

2 1 1 1 2 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

1 1 1 1 1 1 1 1 1 1 3 2 2 2 1 1 1 1 1 1 1 1 1 1

International TWG administration Front End Processes Lithography Interconnect Yield Enhancement Front End Processes Factory Integration Overall Roadmap Technology Characteristics

4 2 2 2 1 1 1 1

2 1 1 3 1 1 1

96

37

TABLE II: Illustrative data on the stretching practice of roadmapping with regard to CMOS and Beyond CMOS. Roadmapping CMOS

Data source Archival data

Illustrative evidence “In the last three decades, the growing size of the required investments has motivated industry collaboration and spawned many R&D partnerships, consortia, and other cooperative ventures […] The overall objective of the ITRS is to present industry-wide consensus on the “best current estimate” of the industry’s research and development needs out to a 15-year horizon. As such, it provides a guide to the efforts of companies, universities, governments, and other research providers or funders. The ITRS has improved the quality of R&D investment decisions made at all levels and has helped channel research efforts to areas that most need research breakthroughs” (ITRS, 2012: 1)

Conference data Observation: the presenters relate to previous milestones set by the ITRS as offering "the path" that needs to be pursued to continue Moore's law (i.e. scaling CMOS-related technologies). Conference slide: "ITRS has demonstrated value of roadmapping for CMOS - Identify pre-competitive research domains, enabling cooperation between industries, institutes and universities. Sharing of R&D efforts. Reduction of development costs and time" (More than Moore TWG, 2011: 4)

Beyond CMOS

Interview data

“I think the industry roadmap is the most important document because that is the collective view of the direction the industry is moving and what the next steps would be [...] So the roadmap is basically saying “Okay, we are going to scale according to Moore’s law. We are going to print dimensions according to Moore’s law” [...] The industry always assumed that you had a roadmap and you could fund research activities associated with this roadmap” (I-115)

Archival data

“"It is also hoped that, by the end of this decade (2020), it will be possible to augment the capabilities of CMOS by introducing new devices that will realize some “beyond CMOS” capabilities. However, since these new devices may not totally replace CMOS functionality, it is anticipated that either chip-level or package-level integration with CMOS may be implemented” (ITRS, 2012: 2)

Conference data Observation: during a coffee break an impromptu interviewee mentions that he has lobbied for about nine months to launch a new TWG related to the area of Beyond CMOS. Towards this end he was collaborating with various colleagues having similar interests and they formulated an internal presentation and written document that highlights the value of the targeted new TWG. Conference slide: key messages from the Emerging Research Materials TWG are, among others, the identification of "New TWG Requirements for ERM […] Transitioning Mature Materials to TWGs […] Workshops" (Emerging Research Materials, 2011: 3) Interview data

“So, they [the future members of a TWG regarding Beyond CMOS] did a first draft and […] convinced the others [ITRS chairs and members] to say: "Ok, this is worth it to have a working group on your own and a Chapter [that is, TWG] on your own and a roadmap." This is a type of thing that would go on” (I-140)

38

TABLE III: Illustrative data on technological uncertainty for CMOS. Uncertainty Technologyrelated

Data source Archival data

Illustrative evidence Presentation slide: Moore’s Law is there to set common goals. The power of Moore’s Law observation and prediction was, is and will be for the foreseeable future to provide a common, easily understood quantified metric for everyone in the semiconductor and IT economy to synchronize their efforts toward historically based, well defined, sustainable and mutually rewarding growth goals in the future. Roadmaps are there to debate the path. The path to achieving those goals (roadmap) was, is and will be subject to unending debates, as it reflects the fundamental uncertainty of assessing risks to schedule and yields of ever more complex novel technologies over extending existing “tried and true” approaches beyond their originally defined limits in the absence of data (Borodovsky 2006, 31; SPIE conference in San Jose, U.S.A.)

Conference data Observation and audio data: conference participants constantly relate to the previous milestones on the ITRS as well as future ones and relate to Moore's law, to which they have geared their collaborative activities in the past, present and will continue doing so in the future (2009-10-20). Interview data

"We know as of today that the CMOS world, as we know it, certainly will continue to exist for a couple of [technological] generations. That common sense across experts; at least regarding scaling in line with Moore's law from a physical perspective there are no challenges we are not aware of" (I138)

39

TABLE IV: Illustrative data on uncertainty for Beyond CMOS. Uncertainty Technology-related

Data source Archival data

Illustrative evidence "A successful new information-processing paradigm most likely will require a new platform technology embodying a fabric of interconnected primitive logic cells, perhaps in three dimensions." (Hutchby, 2002: 39)

Conference data Observation: the organizations admit during a session in a technical working group devoted to Beyond CMOS that they are trying to identify technological candidates that might represent the next technological breakthrough for the Beyond CMOS arena. However, as of yet they do not know which technological paths to follow.

Partner-related

Interview data

"Several [technological] paths are being explored. So it is of interest for all the players in the industry to try to replicate the success we had with Moore's law" (I-130)

Archival data

"In considering the many disparate new approaches proposed to provide order of magnitude scaling of information processing beyond that attainable with ultimately scaled CMOS, the Emerging Research Devices Working Group proposes the following comprehensive set of guiding principles. We believe these “Guiding Principles” provide a useful structure for directing research on any “Beyond CMOS” information processing technology to dramatically enhance scaling of functional density and performance while simultaneously reducing the energy dissipated per functional operation. Further this new technology would need to be realizable using a highly manufacturable fabrication process" (ITRS 2009, 41)

Conference data Observation: at a discussion across technical working groups, it is debated which technologies and accompanying organizations might become relevant in the future. As of yet it is deemed unclear, which kind of organizations to engage with.

Roadmapping-related

Interview data

"We did a series of workshops in the energy domain ... in several areas, but the ones that were most fruitful were the energy domain and bioelectronics. For bioelectronics, the first round table that we had, we brought together colleagues from academia, industry and government from around the world on the electronics side and on the bio side. Just to see if there are points of convergence, common language, points of interest" (I121)

Archival data

“Looking back, we can see how the research of the ‘90s has enabled the breakthroughs of the past decade […] making it essential to plow the seeds of research for the coming decade [regarding Beyond CMOS) into the past one. For this purpose, two additional ITWGs were added, one addressing Emerging Research Devices (ERD) and one addressing Emerging Research Materials (ERM)” (Gargini 2012, 3) in Future Fab Intl. 2012, issue 40, pp. 2-3 “Special Introduction”

Conference data Observation: in a separate workshop integrated into the overall annual ITRS spring meeting in 2011, selected organizational members meet to exchange ideas on how to develop several separate roadmaps for different customer segments. The most likely candidates are the health and automotive sectors. Interview data

"it still possible to develop roadmaps with an s because there will be many of them [needed for Beyond CMOS]" (I-130)

40

FIGURE 1: Development of technological options over time.

Legend: EUV= Extreme Ultraviolet Lithography RF = Radio Frequency  HV = High Voltage QCA =  Quantum Celluar Automata 

Carbon‐based Nano‐electronics Molecular Devices QCA

Spin‐state Transistors Capacitance‐based Memory

Beyond CMOS paradigm

Resistance‐based Memory

Unconventional/Non‐standard CMOS 193 nm Optical Lithography & Enhancements EUV Lithography & Enhancements

CMOS paradigm

PXL EPL PEL IPL

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

t

41

FIGURE 2: A framework for collaborative stretching practices in the face of technological discontinuities.

Practices used with regard to an as of yet unknown technological paradigm

Practices used with regard to an existing technological paradigm

 objective: exploiting technological options within an existing technological paradigm  uncertainty: high, in the main technological in nature  composition of the organizational field: known  collaborative practices: exploited

- signification - legitimation - domination

 objective: search for a new technological paradigm  uncertainty: extremely high, not only of technological nature, but also regarding possible partners and procedures  composition of the org. field: unknown  collaborative practices: challenged, to be refined or possibly substituted

42