You are on page 1of 59

Bachelor Thesis

The Entrepreneurial Ecosystem


of Edge Computing

Chair of Entrepreneurship, Innovation & Technological Transformation

First supervisor: Prof. Dr. Dries Faems

Second supervisor: Maxim Mommerency

Vallendar, May 20th, 2020

Caroline Vaessen

Student ID: 20001680

Date of Birth, Place of Birth: 03.01.1998, Bruges

Address: Grubenstraße 26, 53179 Bonn


Abstract

New use cases, such as autonomous driving, require the processing of huge amounts of real-time data and

the protection of the user's privacy. Centralized solutions, like cloud computing, cannot provide sufficient

reliability in such critical areas. Hence, applications need to move closer to the edge of the data source, which

can be enabled through a local edge server.

This research paper qualitatively analysed the entrepreneurial ecosystem of edge computing using semi-

structured interviews, data from Crunchbase, and complementary research. The entrepreneurial ecosystem

of edge computing consists of independent players and factors that vary from region to region. Differences

between Europe and the USA were identified through the analysis. Compared to the US, European customers

are: Less willing to try out new technologies; there is no sophisticated support system in place; financial

investors are much more conservative. If Europe intends to lead the edge computing market in the future,

policymakers should support small niche providers and practitioners should be aware that some business

cases are only able to a limited extend, like in IoT.

Keywords: Entrepreneurial Ecosystem, Edge Computing, Start-Up, Cloud, Autonomous Driving

I
Table of Contents

List of Abbreviations ................................................................................................................................................. IV


1. Introduction .............................................................................................................................................................. 1
2. Literature Review ..................................................................................................................................................... 3
2.1. Literature Review on Entrepreneurial Ecosystems .............................................................................................. 3
2.1.1. Core Definitions of Entrepreneurial Ecosystems.......................................................................................... 3
2.1.2. Core Challenges of the EE Approach ......................................................................................................... 11
2.1.3. Core Solutions with a Perspective on the Core Challenges ........................................................................ 12
2.2. Literature on Edge Computing .......................................................................................................................... 13
2.2.1. Definition of Edge Computing ................................................................................................................... 14
2.2.2. Core Challenges of Edge Computing ......................................................................................................... 17
2.2.3. Core Solutions of Edge Computing ............................................................................................................ 17
3. Methodology ............................................................................................................................................................ 19
4. Findings ................................................................................................................................................................... 21
4.1. Edge Computing and its Related Technology Fields......................................................................................... 21
4.2. Opportunities and Challenges of Edge Computing ........................................................................................... 24
4.2.1. Key Opportunities of Edge Computing ...................................................................................................... 24
4.2.2. Key Challenges of Edge Computing .......................................................................................................... 26
4.3. The Edge Computing Market ............................................................................................................................ 29
4.3.1. Market Growth and Maturity ...................................................................................................................... 29
4.3.2. Key Players on the Market ......................................................................................................................... 32
4.4. The Funding Landscape of Edge Computing Ecosystems ................................................................................ 33
5. Discussion ................................................................................................................................................................ 36
5.1. Summary of the Findings................................................................................................................................... 36
5.2. Implications for Policymakers and Practitioners ............................................................................................... 38
5.2.1. Implications for Policymakers .................................................................................................................... 38
5.2.2. Implications for Practitioners...................................................................................................................... 39
5.3. Limitations of the Thesis and Opportunities for Further Research.................................................................... 42
References..................................................................................................................................................................... V
Appendix....................................................................................................................................................................... X
Appendix 1: Edge Computing Firms in Europe ....................................................................................................... X
Appendix 2: Edge Computing Firms in the United States ...................................................................................... XI
Appendix 3: Automation Pyramid........................................................................................................................ XIII
Appendix 4: Elements of an autonomous driving system ....................................................................................XIV
Declaration of Authorship ....................................................................................................................................... XV

II
Table of Figures

Figure 1 Entrepreneurial Ecosystem Pillars and their Components ............................................................ 6


Figure 2 The Disruptive Innovation Model .................................................................................................. 10
Figure 3 Key elements, outputs, and outcomes of the entrepreneurial ecosystem .................................. 13
Figure 4 Dynamics that Unfold During the Introduction of Disruptive Innovation ............................... 13
Figure 5 Akamai Technologies Client Request ........................................................................................... 15
Figure 6 Simplified Graphic of Digital Manufacturing .............................................................................. 18
Figure 7 List of Interviewees .......................................................................................................................... 20
Figure 8 Main Challenges and Limitations of the Cloud............................................................................ 23
Figure 9 Key Opportunities and Challenges of Edge Computing ............................................................. 24
Figure 10 Autonomous Vehicles Rely on Several Main Sensor (Sub)Systems ........................................ 25
Figure 11 The Product Life Cycle.................................................................................................................... 30
Figure 12 Global Edge Computing Market .................................................................................................... 30
Figure 13 Edge Computing Start-Ups in Europe ........................................................................................... 31
Figure 14 Edge Computing Start-Ups in the United States .......................................................................... 31
Figure 15 Key factors within the entrepreneurial ecosystem of edge computing (by region) ................. 37

III
List of Abbreviations

AI Artificial Intelligence

AR Artificial Reality

CCPA California Consumer Privacy Act

CRM Consumer Relationship Management

EE Entrepreneurial Ecosystem

GDPR General Data Protection Regulation

IIoT Industrial Internet of Things

IoT Internet of Things

IT Information Technology

OEM Original Equipment Manufacturer

RIS Regional Innovation Systems

Unicorn Start-up valued at more than $1 billion

VR Virtual Technology

IV
1. Introduction

This research paper deals with the entrepreneurial ecosystem of edge computing. New technological

solutions, like edge computing, regularly offer investment opportunities to entrepreneurs, financial investors,

or established companies. The entrepreneurial ecosystem of such a technology consists of independent

organizations that interact with each other. However, an entrepreneurial ecosystem also depends on several

factors within a given region, such as market accessibility, the legal framework, culture, capital.

New use cases, like autonomous driving, require the processing of vast amounts of real-time data, intelligent

data management, and the protection of privacy. Centralized solutions, as cloud computing, cannot provide

sufficient reliability in critical areas. Hence, applications need to move closer to the edge of the data source.

This can be enabled by means of a local edge server.

What information is available on entrepreneurial ecosystems or in the field of edge computing? What are the

specific capabilities of edge computing? How do the market and the financing landscape in Europe compare

to the US? What are useful implications for policymakers and practitioners concerning edge computing? The

paper will try to answer these questions throughout a qualitative analysis.

This paper will consist of five main parts: The Introduction, the Literature Review, the Methodology, the

Findings, and the Discussion. First, the thesis delivers a coherent literature review on the Entrepreneurial

Ecosystems (EE) approach and edge computing. The next part explains the methodology of the qualitative

analysis. The subsequent section consists of findings from semi-structured interviews, the data collected on

Crunchbase, and complementary research. Finally, the discussion part summarizes the core findings,

examines necessary implications, and shows the limitations of the previous analysis.

What are the core findings of this paper? The definition of edge computing depends on the perspective and

application under consideration. Two dominant views are the telecommunication perspective and the IT

perspective. Key opportunities for the technology include (1) the ability to overcome the limitations of the

central cloud, (2) enabling real-time data management independent of the traditional telecommunications

1
network or cloud, and (3) protecting privacy. The main challenges are (1) the misleading definition of the

term, (2) the difficulty of securely connecting multiple devices, (3) handling data without an underlying

managed infrastructure, and (4) country-specific regulations. Currently, the edge computing market is

growing and attracts an increasing number of established companies, especially in the United States, Asia

Pacific, and Europe. At the moment, however, there is no dominant edge computing provider in any of those

regions. The main drivers behind the growth are new use cases, like 5G, AI, AR, or VR. If Europe wants to

be the future leader in edge computing, it needs more capital, the acceptance of failure, and a supportive

environment.

Overall, the findings provide many practical implications for policymakers and practitioners. In summary,

politicians should not exclusively focus on high-performing companies, as several niche players are solely

operating withing specific areas of application. Thus, in Europe, it would make sense to facilitate multi-party

cooperation. One example is the development of a decentral tracing-app for Covid-19 suspects. On the one

hand, practitioners need to be aware that edge computing business cases will not always stack up, especially

concerning IoT. On the other hand, there appears to be an endless amount of investment opportunities.

Notably, emerging markets seem to offer the potential to implement edge computing solutions.

2
2. Literature Review

The following literature review provides an overview of relevant information collected in the past. Hence,

the first part of this section aims to present the most relevant literature available on entrepreneurial

ecosystems, defines the approaches and ideas associated with it, and highlights its key challenges and

solutions. The second part defines edge computing and identifies the technology's challenges and solutions.

2.1. Literature Review on Entrepreneurial Ecosystems

In recent years the subject of entrepreneurial ecosystems has increased in interest (Acs et al., 2017).

Entrepreneurial ecosystems are defined as “a set of interdependent actors and factors coordinated in such a

way that they enable productive entrepreneurship” (Stam, 2015, p. 1765). Due to incrementing importance,

the EE approach has emerged from several theories. The following two paragraphs mention two dominant

lineages, namely the regional development literature and the strategy literature. Afterwards, the EE approach

is defined in greater detail. Digital entrepreneurship and disruptive technology are also explained. As a result

of the sections mentioned above, the core challenges and core solutions are presented.

2.1.1. Core Definitions of Entrepreneurial Ecosystems

As mentioned above, the following part intends to provide an overview of the most important definitions

related to entrepreneurial ecosystems. These key terms – Regional Development Literature, Strategy

Literature, the Digital Entrepreneurship Ecosystem approach, as well as the Disruptive Technology

Literature - will help to understand the idea behind the EE approach.

2.1.1.1. Regional Development Literature

Overall, the regional development literature aims to explain differences in local performance. Related

theories of the EE approach are the Industrial District approach, the Cluster approach, and the Regional

Innovation Systems (RIS), all sharing a focus on regional performance (Acs et al., 2017; Stam and Spigel,

2017; Teriesen et al., 2017).

3
Marshall (1920) states that the Industrial District approach focusses on the local division of labour in the

industry, and Becattini (1990) describes the firm population in a socio-territorial entity as an essential factor

for the international success of industries. Another related theory, the Cluster approach, which has been

mainly driven by Michael Porter, builds upon the industrial organization literature and analyses related

industries in a specific location. Furthermore, it focusses on the geographic concentration of firms, how

companies are connected, and how they compete (Acs et al., 2017; Porter, 1990; 1998). Isenberg (2010)

links Porter´s Cluster approach with the concept of Entrepreneurial Ecosystems. On the other hand, RIS

addresses the knowledge spill-over effect between knowledge-producing institutions and innovative firms

within a particular region. This exchange is capable of increasing the overall innovativeness of an area

(Cooke at al., 1997). Knowledge hubs are, for example, Universities or other public research facilities (Acs

et al., 2017).

2.1.1.2. Strategy Literature

In contrast to the regional development literature, the strategy literature takes into account that ecosystems

need to be perceived in a global context (Zahra and Nambisan, 2011). Moreover, the strategy literature

analyses the value creation, and value capture of individual companies and assumes that the leadership in an

ecosystem is taken over by a focal company (Acs et al., 2017). On the one hand, the available literature

assumes that the firm´s ability to capture value depends on structural and strategic factors (Adner and

Kapoor, 2010; Jacobides et al., 2006). On the other hand, the value creation critically depends on different

actors within a business ecosystem (Iansiti and Levien, 2004; Adner and Kapoor, 2010; Williamson and De

Meyer, 2012). Adner et al. (2013) claim that a company's strategy must be attractive not only to end-

consumers but also to other stakeholders, like business partners. In conclusion, this implies that well-

managed and networked ecosystems create value for everyone involved, as well as that interaction or

interdependencies among all stakeholders, need to be well managed (Uzunca et al., 2016)

Besides, there are related streams of strategy literature, emphasizing the role of platforms, surrounding an

ecosystem, and connecting customer groups (Acs et a. 2017; Rochet and Tirole 2003; Parker and Van

4
Alstyne 2005; Evans and Schmalensee 2016). Typical platforms are Facebook or AirBnB and are described

in the Digital EE approach. However, Autio and Thomas (2014) claim that platforms do not necessarily play

a central role in ecosystems and are therefore not a defining feature of them.

2.1.1.3. The Entrepreneurial Ecosystem Approach

The two dominant lineages, the regional development literature and the strategy literature, are essential to

understand the idea behind the EE approach. Before differences and similarities between the regional

development literature and the strategy literature will be described, there is a need to identify several

independent terms of the EE approach.

One relevant term is entrepreneurship. Stam (2015) and Schumpeter (1934), define entrepreneurial activity

as the process through which individuals create or exploit opportunities for innovation. Innovation is the

optimal outcome of entrepreneurial ecosystems. Moreover, entrepreneurship is something that takes place

in a community of many independent and different parties (Stam, 2015). Different parties are, for example,

policymakers, firms, end-customers, or financial investors. Hence, entrepreneurship is also capable of

creating positive socio-economic effects (Acs et al., 2017). However, it is crucial to note that not all new

firms can develop innovative products and services or profit as a result of economic growth and employment

(Davis et al. 1996). In the past, entrepreneurship studies failed to link the role of systems “in explaining the

prevalence and performance of entrepreneurship” (Acs et al., 2017, p. 1). A system is a “set of interacting

and independent organizations that function together as a whole to achieve a purpose” (Sussan and Acs,

2017, p. 57). Ecosystems are defined as interactions between living and non-living components (Acs et al.,

2017; Tansley, 1935) and can be seen as a network of dynamic and interacting systems that depend on several

factors in their respective context (Sussan and Acs, 2017). Many researchers define attributes, principles, or

other factors that need to be considered to establish an entrepreneurial ecosystem. According to Isenberg

(2011), there are six distinct domains in ecosystems, namely policy, finance, support systems, human capital,

culture, and markets. According to the World Economic Forum (2013), there are not six but eight pillars

(Figure 1), which are essential to consider in EE. Those pillars are Accessible Markets, Human Capital,

5
Financing, Support Systems, Regulation, Education, Major Universities as Catalysts, and Cultural Support.

The pillars focus on the critical resources in ecosystems. Mentioning factors or distinct domains in

ecosystems helps to understand relevant contributors to a thriving ecosystem. However, defining such

elements often ignores underlying causes or unique attributes of a region to the success of entrepreneurial

ecosystems. Similar to economics, ecosystems are about performance. Economic studies explain output,

economic behaviour, and outcomes, aggregate welfare of systems. However, economic studies ignore the

role of entrepreneurship in ecosystems (Acs et al., 2017; Sussan and Acs, 2017). According to Sussan and

Acs (2017), entrepreneurship is an economic output, constrained by its context. Terjesen et al. (2017)

suggest that economic performance is the ultimate motivation for entrepreneurial ecosystems. According to

the researchers, economic performance is not limited to individuals, companies, or regions. Furthermore,

economic regions are supposed to devote considerable effort and resources to achieving and maintaining

strong performance (Terjesen et al., 2017).

Figure 1: Entrepreneurial Ecosystem Pillars and their Components

Pillar Components
Accessible Domestic market: large/ medium/ small companies as customers and
markets governments as customers
Foreign market: large/ medium/ small companies as customers and
governments as customers
Human Capital/ Management talent, technical talent, entrepreneurial company experience,
Workforce outsourcing availability and access to immigrant workforce
Funding & Friends and Family, angel investors, private equity, venture capital and access
Financing to debt
Support Systems/ Mentors/ advisors, professional services, incubators, accellerators and
Mentors networks of entrepreneurial peers
Government & Regulatory Ease of starting a business, tax incentives, business friendly legislation/
Framework policies, access to basic infrastructure, access to telecommunications/
broadband and access to transport
Education & Available workforce with pre-university education, available workforce with
Training university education and those with entrepreneurship-specific training

Major Universities as Promoting a culture of respect for entrepreneurship, playing a key role in idea-
Catalysts formation for new companies and playing a key role in providing graduates to
new companies
Cultural Tolerance for risk and failure, preference for self-employment, success stories/
Support role models, research culture, positive image of entrepreneurship and
celebration of innovation

Source: World Economic Forum (2013, pp. 6-7)

After the identification of some independent terms of the EE approach, the next part highlights the

differences to the related approaches. On the one hand, the EE approach is similar to the strategy and the

regional development literature, as they all stress the independence between value-creating actors and factors

in a specific community (Acs et al., 2017). Moreover, all approaches have a focus on the external business

6
environment (Stam, 2015). Also, the focus on aggregate value creation within a specific region is similar

between EE literature and the Regional Development approach (Acs et al., 2017).

However, the EE approach clearly distinguishes itself from other theories. The concept focusses on the

individual entrepreneur as a value creator and focal point in an ecosystem (Acs et al., 2017; Pitelis 2012;

Stam and Spigel 2017), and therefore, has an interest in decreasing the role of the government (Stam, 2015).

Nevertheless, only ambitious entrepreneurs aim to add as much value as possible to the society, can achieve

economic growth, and only this small group of entrepreneurs can be interpreted as the basis of Schumpeter’s

welfare theory (Stam et al., 2012; Stam, 2015; Schumpeter, 1934). Even if the EE approach starts with the

entrepreneurial actor, it is essential to realize this emphasis on productive entrepreneurship (Stam, 2015).

Due to the growing recognition of this particular type of entrepreneurial actors, the attention of politicians

has shifted from increasing the quantity of entrepreneurship to improving its quality (Stam, 2015). Stam

(2015) further claims a transition from entrepreneurial policy to policies for entrepreneurial ecosystems.

Moreover, the EE approach sees entrepreneurship as its output, playing an essential role in creating the

ecosystem as well as in creating employment, innovation and growth in society (Acs and Audretsch, 1988;

Acs et al., 2017, Davis et al. 1996; Stam, 2015). By setting a different focus, compared to other related

theories, the EE theory can explain the rise of Unicorns, “start-ups valued at more than $1 billion” (Acs et

al., 2017, p. 6). According to Acs et al. (2016) and Evans and Schmalensee (2016), those Unicorns can build

a highly scalable platform where value is being created and exchanged among its actors. Besides,

entrepreneurs can detect market inefficiencies (Sussan and Acs, 2017). Even if government actions are

insufficiently taken to eliminate system or market failures, the ambitious entrepreneur can find opportunities

in creating innovative goods and services (Stam, 2015). In comparison, the regional development literature,

for example, does not recognize any central leadership role beyond the facilitating role of a government

(Amin and Thrift, 1994; Acs at al. 2017). Moreover, in the EE approach, the entrepreneur is likely to partner

up with the public sector (Acs et al., 2017). Stam (2015) and Teece (1992) emphasize that non-market

interactions are equally important to realize innovation as market interaction because not all factors in

7
innovation systems can be reduced to the market context. Also, cooperation between all actors facilitates an

increase in knowledge sharing (Stam, 2015).

Besides looking at the individual entrepreneur, the EE approach observes the city, regional, and national

contexts (Acs et al., 2017). According to Anselin et al. (1997) and Alvedalen and Boschma (2017), cities

deliver knowledge and human capital, whereas knowledge spill-overs happen locally, despite globalization.

On the other hand, Bruns et al. (2017) covered 107 European regions in 16 European countries to search for

the existence of entrepreneurial ecosystems. They concluded that ecosystems are impossible to measure

directly and detected no evidence of heterogeneous entrepreneurial activity across regions. Moreover, Acs

et al. (2017) were unable to identify evidence that entrepreneurial ecosystems impact performance at the

country level.

2.1.1.4. The Digital Entrepreneurship Ecosystem Approach

Concerning the EE approach, it is worth taking a look at a specific area, namely that of Digital

Entrepreneurship Ecosystems. The terminology of digital ecosystems emerged in the early 2000s (Sussan

and Acs, 2017). A digital ecosystem is defined as a “self-organizing, scalable and sustainable system

composed of heterogeneous digital entities and their interrelations focusing on interactions among entities

to increase system utility, gain benefits, and promote information sharing, inner and inter cooperation and

system innovation” (Li et al. 2012, p. 119). The framework integrates the idea of digital ecosystems into the

EE approach (Sussan and Acs, 2017). Sussan and Acs (2017) observe two foundational pillars, namely

digital technologies and users. Moreover, this specific type of ecosystem assumes a bottom-up approach that

is driven by its users and mainly open source-oriented (Dini et al., 2011). Those open digital systems enable

every user to participate, become entrepreneurs, and enable innovation with little limitations in place (Sussan

and Acs, 2017). The application of digital ecosystems is possible; for example, in business, knowledge

management, education, or social networks (Sussan and Acs, 2017). Facebook, Tripadvisor, or AirBnB have

created successful examples of digital ecosystems, as they invented multisided platforms that depend, for

example, on the content, or tangible assets of its users (Sussan and Acs, 2017; Richter et al., 2015). Moreover,

8
these companies can match several groups of customers, meanwhile reducing the transaction costs of a match

(Coase, 1937). Overall, a shift towards more human interactions in digital technologies can be observed,

creating the need for inclusive, dynamic, and flexible digital infrastructures to capture digitization effects

(Sussan and Acs, 2017; Tilson et al., 2010).

2.1.1.5. Disruptive Technology Literature

After the discussion of related theoretical approaches of the EE, it is worth taking a look at the terminology

of disruptive innovation, or, respectively, disruptive technology. One can think about it from the perspective

of Clayton M. Christensen, who has dominated the general understanding of the term among academics over

the last 25 years. According to Christensen and Bower (1996), disruptive products or services encroach from

the low-end of the market upwards. At the beginning of the development stage, it “cannot be used by

customers in the mainstream markets” (Christensen et al., 2004, p. 293), because it has inferior product

characteristics, such as capacity. Over time, disruptive innovations improve to the point where they can

compete with the old product or service attributes of the incumbents (Figure 2). They do this by delivering

superior new product features. According to Christensen and Bower (1995; 1996), disruptors are typically

small and agile firms capable of changing product and market strategies. As soon as they established their

disruptive product or service on the market, they continuously improve and attack the established business

of incumbents from the low-end of the market. One reason why many established companies fail to react to

disruption is that they are usually only able to maintain their leading position within particular technology

trajectories. In contrast to established firms, small companies can change their strategies quickly, are more

likely to identify and exploit technological opportunities, and are not, to a certain extent, constrained by past

expertise and history (Bower and Christensen, 1995; Charitou and Markides, 2003). One recent example of

disruption is Netflix, taking advantage of the internet to stream videos and taking market share of

Blockbuster (Christensen et al. 2015).

9
Figure 2: The Disruptive Innovation Model
Higher

Product
Performance

Lower Time

Source: Christensen et al. (2015, p. 7)

In addition to Christensen’s perspective, other researchers are continuing the idea in different areas. It should

be noted that Christensen only considers linear business models and does not include platforms and

ecosystems in his work (Moazed and Johnson, 2016). Regarding the recent finding of Ansari et al. (2015),

disruptors confront the disruptor’s dilemma when introducing their innovation into multisided ecosystems.

In contrast to Clayton M. Christensen, Ansari et al. (2015) emphasize not only the disruptive effect on

incumbents but also the effect on an entire ecosystem. Ansari et al. (2015) focus on the possibility for start-

ups to gain the support of precisely those established companies that they disrupt, observing the simultaneous

presence of cooperation and competition (Ansari et al., 2015; Bengtsson and Kock, 2000; Brandenburger

and Nalebuff, 1996; Gnyawali and Park, 2011; Ketchen et al., 2004). Overall, the research focuses on the

opportunity for small disruptive firms, rather than on challenges. To illustrate their findings, Ansari et al.

(2015) mention the example of TiVo in the U.S. television (TV) ecosystem. TiVo entered the TV industry

by connecting all stakeholders on its multisided platform. For example, TV viewers, content providers, or

advertisers. The initial reaction of incumbents was rather adverse, as the benefits were only visible in the

long run, but the disruptive forces were immediately felt. Thus, TiVo continuously adapted its strategy and

reacted to feedback whenever it was required while keeping its vision in mind. One measure was hiring an

executive familiar with the media industry to facilitate communication with different actors. Another

measure was the decision to form collaborative ventures with incumbents. Whenever tensions could not be

resolved, TiVo cooperated only to the possible extent and decided to let the tension remain. Through constant

adaptation and dynamic adjustments to the strategy, the advantages of TiVo became visible, and the support

increased accordingly.

10
2.1.2. Core Challenges of the EE Approach

Comprehensive research on entrepreneurial ecosystems is only about five to ten years old and is only now

attracting increasing attention on the government side. The EE approach shifts the traditional economic

thinking about businesses, markets, and market failure, towards a new perception of individuals,

organizations, and networks (Stam, 2015). However, are functional entrepreneurial ecosystems leading to

successful entrepreneurship or, the other way around, is successful entrepreneurship simultaneously

guaranteeing entrepreneurial ecosystems? Stam (2015) questions this tautological reasoning. The EE

approach still offers no universal definition, and the critical success factors of entrepreneurial ecosystems

are only superficially described. “[…] the rush to employ the EE approach has run ahead of answering many

fundamental conceptual, theoretical, and empirical questions” (Stam, 2015, p.1763). Furthermore, the

concept focuses exclusively on high-growth start-ups, which could lead to a too exclusive view (Stam, 2015).

The reason for this focus is the observation that only those firms are essential sources of innovation,

economic growth, and employment (Mason and Brown, 2014; World Economic Forum, 2013). However,

according to Stam (2015) and Baumol (1990), small businesses or entrepreneurial employees have been left

out, even if they can be forms of productive entrepreneurship and, thus, welfare outcomes.

Small, disruptive firms need to own or need to be able to get access to resources, like financing or skilled

personnel, to become successful (Ansari et al., 2015). Usually, start-ups lack resources, such as financing,

leading to a low survival rate (O'Reilly and Tushman, 2011; Stubbart and Knight, 2006). Thus, start-ups are

forced to cooperate with other firms and build a value-added network to establish their innovation

(Barndenburger and Nalebuff, 1996; Teece, 1986). Cooperation with different firms can be a considerable

challenge, as essential partners might be the ones that are being disrupted (Ansari et al., 2015). Gilbert and

Bower (2002) suggest acquisitions as a possibility for incumbents to be successful with disruptive

technologies. Thereby, start-up firms can, however, not become leaders of ecosystems, as described by

Ansari et al. (2015).

11
When looking at digital ecosystems, like the one of TiVo, another problem arises. Multisided platforms are

difficult to govern as they combine many different actors with various interests and needs (Acs et al., 2017;

Sussan and Acs, 2017). TiVo, for example, simultaneously needs to cope with TV viewers, advertisers,

content providers, or regulators. TiVo, which was confronted with a complex ecosystem, managed the

situation well. However, this might be an exception. Would cooperation be realistic for the majority of start-

ups? Only a few companies can cope with many diverging actors. There is considerable uncertainty to attract

a sufficient number of adopters to build a multisided platform (Caillaud and Jullien, 2003; Evans et al., 2006;

Rochet and Tirole, 2003). What happens if disruptors are not able to meet future expectations, and what if

they fail to manage disappointment (Ansari et al., 2015)? Mismanagement can, on the other hand, lead to

the loss of legitimacy with ecosystem members and thus to even more severe consequences (Ansari et al.,

2015).

2.1.3. Core Solutions with a Perspective on the Core Challenges

How can the problem of superficially defining key success factors of ecosystems be solved? Stam (2015)

elaborated on four ontological levels, including the relationships between them. Furthermore, his graphical

illustration (Figure 3) makes a distinction between framework and systematic conditions. Overall, the figure

helps to understand critical elements as well as outputs and outcomes of ecosystems. Stam (2015) and Thurik

et al. (2013) further demand policy for an entrepreneurial regional economy, to create a context where

productive entrepreneurship can grow. Thus, the policy focus would no longer limit itself to maximizing one

specific factor. Besides looking at inclusive institutions like TiVo, one could think of the entrepreneurial

ecosystem as open systems with shared responsibility (Stam, 2015). Again, this would require supportive

political institutions (Acs et al., 2017). If this perspective is adopted, the focus on high-growth start-ups can

be expanded to include small businesses and entrepreneurial employees as forms of productive

entrepreneurship.

12
Figure 3: Key elements, outputs, and outcomes of the entrepreneurial ecosystem

Outcomes Aggregate Value Creation

Outputs Entrepreneurial Ecosystem Elements

Entrepreneurial Activity
Systemic Support services/
Networks Leadership Finance Talent Knowledge
conditions intermediaries

Framework Formal Physical


Culture Demand
conditions institutions infrastructure

Source: Stam (2015, p. 1765)

To facilitate the government of multisided platforms, Ansari et al. (2015) introduced a process model (Figure

4). It is a process model of dynamics that unfold during the introduction of disruptive innovation. The figure

illustrates how, on the one hand, disruptors can break existing ecosystems apart, but on the other hand, they

can also resolve tensions within these systems (Ansari et al., 2015; Nye, 2004).

Figure 4: Dynamics that Unfold During the Introduction of Disruptive Innovation


Platform Transformation

Coopetitive Spill overs &


Firm s Actions
Tensions Accommodations
Provide
Intertemporal vision of
coopetition future
Accommodating
benefits
one incumbent or
side spill over to
others
Disruptive Sustaining Modified
Innovation Frame Build critical Frame Innovation
Dyadic
coopetition mass of
DISRUPTS consumers SUSTAINS
Firm
continues to
Existing adjust its Changed
Ecosystem strategy and Ecosystem
Multilateral Simultaneously technology
coopetition engage multiple
incumbents

Ecosystem Transformation

Source: Ansari et al. (2015, p. 1846)

2.2. Literature on Edge Computing

After having reviewed the available literature on entrepreneurial ecosystems, the next part introduces the

specific topic of edge computing. The term will be defined first; afterwards, core challenges and solutions

will be presented. Overall, this section provides a rough overview of the technology. The topic will be

analysed in more detail in the findings section and the concluding discussion.

13
2.2.1. Definition of Edge Computing

Edge computing evolved from different areas. One dominant area is Information Technology (IT). Another

area is that of telecommunications. The general idea of edge computing is to bring applications closer to the

data sources (IBM, 2020), where all the data processing happens in real-time on a local edge server. In other

words, edge computing is serving as a “decentralized extension of the campus networks, cellular networks,

data centre networks or the cloud” (van der Meulen, 2018).

When looking at the IT perspective, edge computing is everything that happens outside the cloud (Reynolds,

2019). Cloud computing operates on big data, whereas edge computing operates on instant data. Gartner, a

leading research and advisory company, estimates that by 2025, 75 percent of the data available today will

be processed outside traditional data centres or clouds (IBM, 2020). Edge computing aims to deploy and

manage cloud infrastructures in a fully distributed manner. At the same time, this extends the employment

reach. The idea of shifting specific processes closer to the data source is exceptionally relevant in sensitive

areas, for example, in autonomous cars. The data within and outside the vehicle is transferred to a local

server in real-time (van der Meulen, 2018), solving the limitations of the cloud, like latency. The particular

case will be dealt with in more detail in the findings section.

Generally, edge computing can be applied anywhere, be it in autonomous driving cars, in drones, in retail,

or digital manufacturing. One of the first ideas to apply something similar to what edge computing

technology is doing today originated in the late 1990s. A company called Akamai Technologies developed

the concept of Content Delivery Systems to overcome bottlenecks from the internet’s edge (Dilley et al.,

2002). The company observed problems for a website’s scalability, reliability, and performance in serving

content from only one location. To overcome this problem, the firm moved from merely providing web

objects to providing local clustering options at the network edge (Dilley et al. 2002) (Figure 5). Already in

2002, the company identified its customer's need to get better visibility of their current applications.

Subsequently, the company further developed its service and enabled web applications on a globally

14
distributed computing platform (Davis et al., 2004). This enabled response times to end-users anywhere in

the world.

Figure 5: Akamai Technologies Client Request

Top-level DNS
Content provider

5 Internet

Akamai Low-level DNS


server
1
4 2
3
Edge server

Client HHTP content request. Once DNS resolves the


edge server s name (steps 1 and 2), the client request
is issued to the edge server (step 3), which then User
requests content from the appropriate source (step 4
or 5), satisfies the request, and logs its completion.

Source: Dilley et al. (2002, p. 4)

There are several reasons why edge computing has become increasingly important over the last five years.

One reason is the growing adoption of IoT devices. Other reasons are the demand for real-time, automated

decision-making solutions, as well as the constant increase of data volumes and network traffic

(MarketsandMarkets, 2019). The 2019 market research report of MarketsandMarkets forecasts the edge

computing market to grow from 2.8 billion USD in 2017 to 9 billion USD in 2024. The report further states

that the acceptance of edge solutions by large companies is high, due to the broad geographical presence and

customer base. Gartner, highlighted edge computing as one of the Top 10 Strategic Technology Trends for

2018 (Panetta, 2017). Currently, many large corporates are investing in edge computing, like Microsoft,

Amazon, Cisco, Huawei, or IBM.

2.2.2.1. Edge Computing as a Disruptive Technology?

In order to understand the nature and origin of edge computing, it is worthwhile to take a look at its disruptive

power. However, as discussed earlier, is edge computing a typical low-end disruption according to the theory

of Clayton M. Christensen? If edge computing solutions were disruptive in Christensen’s sense, it would

initially have inferior product characteristics. Besides, it would encroach the market from the low-end

upwards, as it would be able to deliver superior new product features at a given point in time.

15
Indeed, edge computing has a few disruptive characteristics. By the time Akamai Technologies introduced

the idea of edge computing, it was not considered necessary by many market players. Many manufacturing

processes were not digitized, and the internet was not yet widespread. Also, the number of connected IoT

devices was low. Moreover, in the early 2000s, the focus was exclusively on the cloud or on massive data

centres where the relevant data was stored. Those were sufficient storage options, as there were fewer data

and less critical areas where, for example, latency is essential. However, over time digitalization became

more and more critical to businesses, governments, and individuals all over the world. In the past years, the

number of connected IoT devices continuously increased and is expected to keep rising in the future. It

became apparent that IoT could not happen in a central cloud. As a result, many large companies are now

investing in edge computing to meet the growing demand in the marketplace.

The majority of the investing firms, like Amazon or Microsoft, are also unbeaten in providing cloud

computing. In the future, firms will still rely on the cloud. Therefore, the new technology does not replace

the existing market. Nevertheless, it could scale it down. As a consequence, companies, like Amazon and

Microsoft, combine cloud and edge, and create hybrid models. This could make them very competitive in

comparison to small and young start-up firms. Nevertheless, those small firms have an advantage as they are

independent, unbiased by previous experience, agile, and focused on a particular area. Especially their

independence and small size makes them more trustworthy, particularly for critical applications.

Apart from the disruptive nature of the technology, it is questionable whether edge computing was inferior

in its early stages, given Christensen’s logic (Figure 2). This question is difficult to answer from the available

literature since edge computing is a vast topic that can be applied everywhere. Moreover, these use cases

differ concerning their requirements, which makes it difficult to imagine a standard at some point. Thus, in

summary, edge computing has many disruptive features. However, it does not quite fit into Christensen’s

core concept, as well as into his performance trajectory chart (Figure 2).

16
2.2.2. Core Challenges of Edge Computing

Edge computing has not yet reached maturity, and there are still many general questions unanswered about

the technology’s capacity, as well as its security. As previously mentioned, edge computing is removing

specific processing and managing tasks from the central cloud or data centre to a local edge server. In

comparison to traditional data centres, edge servers store less data (Chemitiganti, 2019). Thus, areas of

responsibility must be clearly defined for each application scenario. Which data will be analysed in real-

time, and what will be transferred to the cloud or data centres? Another critical point is data analysis. What

happens if the maintenance of servers causes an interruption in the management process (Mattews, 2018)?

Interruption can be fatal in critical areas, like in portable medical devices. Moreover, applying edge

technology in critical areas bears considerable security risks. Santhosh Rao, senior research director at

Gartner, claims that extending one’s “footprint using Edge Computing exponentially increases the surface

area for attacks” (van der Meulen, 2018). Has the technology already been sufficiently tested? Besides, is

Artificial Intelligence, which is needed to process and analyse data, able to react to different scenarios

adequately? Also, some claim that 5G is a critical requirement for some edge computing use cases to function

optimally (Mattews, 2018).

2.2.3. Core Solutions of Edge Computing

There are still many unresolved questions. Some of them can be tackled in the short or medium term, such

as deciding what data has to be managed locally. While other issues seem to be more challenging to handle,

such as ensuring that there is no interruption in the real-time data processing.

To illustrate an ideal world, one can think about Digital Manufacturing. Standardization and orchestration

of components, machines, suppliers, or customers in Digital Manufacturing could increase efficiency,

facilitate information flow, decrease the cost of operating, and increase value for all actors. In Digital

Manufacturing, fabrics are connected internally (Figure 6) but lack efficient data transfer to the outside. In

an optimized world, all relevant payers are connected, the “machine language” is standardized, every single

component is connected to the internet (IoT), and data is being processed in real-time. Every firm would

17
produce optimal quantities and adjust its manufacturing to changes in demand without problems.

Furthermore, errors would be eliminated before they occur due to predictive tools.

Edge computing plays a vital role in optimizing processes and facilitating real-time data processing.

However, are edge computing providers capable of creating value through the construction of multisided

ecosystems? Indeed, edge providers could make use of their interconnecting role between data source and

data centre. Especially in digital manufacturing, multisided platforms could make sense. Those platforms

could not only connect edge providers, OEMs, and data centres; they could also integrate campus network

providers, end-users, the machine produces, or suppliers. Nevertheless, the question remains, which actor

has the legitimacy to take a leading role? Is it a start-up company specializing in edge technology or a

financially secure company that may have operated a data centre or cloud before, using its existing

knowledge and network?

Figure 6: Simplified Graphic of Digital Manufacturing

Source: Own Creation

18
3. Methodology

After having reviewed the literature on entrepreneurial ecosystems, as well as the available information on

edge computing, the following part describes how additional data has been collected through qualitative

analysis. This data aims to deepen further the understanding of the capabilities of edge computing and the

ecosystem that surrounds it.

The data on edge computing has been collected mainly from Crunchbase, as well as from seven semi-

structured interviews. Furthermore, the data has been complemented by secondary research, like market

reports, to underline specific interview statements and illustrate the edge computing market structure.

The first research step was the filtering of data on Crunchbase. Crunchbase is a platform consisting of

approximately 55 million professionals, “including entrepreneurs, investors, market researchers, and

salespeople.” (Crunchbase, 2020). Crunchbase offers the possibility to filter companies on specific

categories. My initial database consisted of companies that (1) focus on edge computing, (2) are located in

Europe, and (3) have been founded after 2010. The output table subsequently contained the names of the

companies, their location, a description of the sector, information on the total amount of financing, and the

date of foundation. After that, I manually standardized and adjusted certain information. In total, I have

compiled a table with 23 companies (Appendix 1). At a later point in time, I collected data on Crunchbase

with the same categories for companies in the United States. The collected data resulted in a table consisting

of 43 companies (Appendix 2). The two tables helped to create a comparison between European and

American start-ups in order to draw conclusions about the ecosystem, precisely the difference in investment.

In a second step, I contacted relevant European founders and executives from my Crunchbase data, as well

as several other companies that I identified during the search. All relevant practitioners were contacted on

LinkedIn, directly via mail, or through a contact form on the respective company website. Afterwards, I

executed semi-structured interviews with seven individuals (Figure 7). In general, semi-structured interviews

allow for a flexible adjustment of the agenda, create room for additional questions and discussions. Each

19
interview was transcribed afterwards, and specific statements were used to complement the line of

argumentation within the findings section.

Figure 7: List of Interviewees

To structure the data in a third step, I highlighted relevant sentences in the transcripts and copied them into

a document for which I had previously created an outline. This facilitated the subsequent evaluation. After

that, I looked for particular connections between the interview output and the Crunchbase data. Besides, I

also inserted complementary data, for example, graphs from market reports, to underline my findings. In

summary.

In summary, based on the semi-structured interviews, as well as on the Crunchbase data, I was able to: Create

an overview about the origin of edge computing; assess the technologies capabilities; illustrate the edge

computing market, as well as its ecosystem; and compare the funding in Europe with the North American

one.

20
4. Findings

After the review of the literature on entrepreneurial ecosystems and the short overview of edge computing,

this chapter intends to delve deeper into edge computing. The findings are based on a comprehensive

qualitative analysis, namely semi-structured interviews, on selected data from Crunchbase, and on

complementary secondary research. The first section discusses related technology areas of edge computing

to describe where the idea originated. The second part presents the key opportunities and challenges to assess

the performance of the technology. The adjacent sections illustrate the market for edge computing, its

maturity, and its key players. The final section distinguishes between financing activities in the field of edge

computing between Europe and the USA. It should be noted that all statements about the difference between

American and European investments are made from a European perspective.

4.1. Edge Computing and its Related Technology Fields

Currently, there is no standardized view on edge computing, partly because the term itself is not intuitive.

Nevertheless, it is crucial to understand the meaning, especially concerning specific applications, to get all

the benefits out of the technology. This section provides an overview of where edge computing came from

and how it evolved. This will help to understand some diverging opinions of practitioners and statements

made by interviewees in the course of this paper.

There are several opinions concerning the question of how edge computing emerged. Some suppose its roots

are in Mobile Edge, and others believe it emerged from fog computing. Those perspectives share the same

idea of bringing applications closer to the data source. However, they differ concerning the underlying logic

and their assumptions.

On the one hand, some practitioners support the position that edge computing emerged out of Mobile Edge

Computing, which itself emerged out of the telecommunications space. The idea is that edge computing is a

local base station, for example, in a grocery shop or in an autonomous driving car, where a lot of data

processing takes place. In addition to this local base station, there may be additional physical devices on

21
which other non-critical processes, like the ventilation system, are running. The local edge server is

programmed by algorithms that are capable of performing complex tasks independently of other workloads

of the application. The telecommunications point of view assumes those processes to be standardized. This

logic results from the network idea since no telecommunications network is possible without standardization.

A network is of crucial importance in the telecommunications space to provide both a telephone connection

or an internet connection. However, not everything above the network needs to be standardized. If everything

were standardized, there would be no more opportunity for innovation. Out of this perspective, edge

computing devices sit on top of a standardized network and support critical applications where a stable

connection is vital. Here, local edge computing is necessary because a standardized fibre optic network has

a physical limitation, namely the infinite speed of light. This limitation creates critical latency problems in

poorly connected areas, for example, in the countryside or emerging countries.

“You cannot have a network without standardization […] But once that network
is there, everything that is going on top of that network doesn't need standardization.”

- Jurgen Hofkens -

Contrary to that, the IT perspective supports the position that edge computing emerged out of fog computing.

Therefore, it can be regarded as a logical extension to the cloud. In general, cloud computing and edge

computing are both able to give elastic compute on demand. However, the main difference between cloud

and edge computing is that it is given to the respective customer locally rather than centrally.

“Edge computing […] has all the benefits of moving a


process away from a main, centralized cloud server.”
- David Shackleton -

The cloud aggregates information from multiple computers, each having its intelligent management system.

This system allows people to manage and virtualize data of several workloads. Similar to the perspective of

telecommunications, the IT perspective faces the problem of latency when moving away from the central

cloud, constraining efficiency at the same time. Besides, it is evident that in traditional cloud environments,

there is a management problem. Edge computing use cases only have one or two computers with various

locations, whereas the cloud has multiple computers for each workload inside one site. This can be illustrated

22
from the perspective of a large enterprise, like Circle K. The company has various locations in the United

States. Besides, it offers several different services in each store, like fuel, carwash, beverages, and freshly

made food. Each service has its computer in place with one workload on it, summing up to a lot of technology

in each store. Thus, the management of one central cloud with a large number of computers in it becomes

difficult to handle. Every time a single computer shuts down, the central office has to send technicians to the

respective location, which in turn costs a lot of time and money. Edge computing allows bringing all

workloads into one place in each store, shifting the data management closer to the actual data source.

“They need a virtualization capability on the edge. And they need a


good management system to be able to manage those from the edge.”

- Brian Buggy -

Apart from the management issue of a central cloud, another problem concerns privacy. As an example, one

could think about an industrial company putting all the confidential customer information in the CRM

platform of Salesforce. Salesforce is an American cloud-based company running their data centers only in

nine places around the globe (Salesforce, 2020). If the confidential data is, for example, transferred to a

location in America, the company would have to rely on the ability of Salesforce to secure the data. After

whistleblower Edwards J. Snowden uncovered comprehensive monitoring programs of the US government,

trust in American institutions seems to be fragile, especially among European companies.

“More and more people are starting to push back on this


model of throwing everything into the cloud.”

- Paul Mundt -

Figure 8: Main Challenges and Limitations of the Cloud

Source: Own Creation

23
4.2. Opportunities and Challenges of Edge Computing

After this broad overview of edge computing, it is essential to demonstrate its unique capabilities. Section

one gives an overview of the key opportunities. The second section highlights specific challenges that must

be addressed as the technology evolves.

Figure 9: Key Opportunities and Challenges of Edge Computing

Source: Own Creation

4.2.1. Key Opportunities of Edge Computing

It is already apparent that traditional solutions such as data centres and the cloud are not sufficient to support

critical applications. Especially the emergence of new IoT use cases requires real-time data processing. The

limitations and problems of the cloud, namely latency, data management, and data security, make it clear

that in a long-term perspective a solution like edge computing is needed.

“Cloud computing […] always imposes a minimum latency of actually


network communication […] So that's always going to be a bottleneck.”

- David Reischer -

One key opportunity of edge computing is its promise to solve the limitations of the cloud, like insufficient

connectivity. This can be observed in the example of autonomous driving. There is never one hundred

percent connectivity when driving a car in rural areas. However, when people rely on a critical function,

missing connectivity can have negative consequences. For example, the autopilot must react immediately if

a child suddenly crosses the road. In order to ensure safety, it is essential that the computing is done in the

vehicle itself and closer to the end-user. Besides, it is a logical step to avoid dependence on a network, to

enable real-time data processing, and to facilitate more intelligent data management at the edge.

24
Figure 10 gives an overview of the essential elements within an autonomous car. Those were investigated in

a study by McKinsey&Company in 2017. Edge computing refers to all real-time data processing that takes

place inside the car. Nevertheless, the cloud plays a vital role in providing traffic data or road maps that

happen outside the car.

Figure 10: Autonomous Vehicles Rely on Several Main Sensor (Sub)Systems

Source: Heineke et al. (2017, p. 5)

Another key opportunity that practitioners observe is the redundancy to construct costly, inefficient, and

insecure fiber-optic networks, especially in countries or continents that lack such an infrastructure. African

countries offer a significant niche for edge computing providers and make not only fiber optics but also the

cloud sustainable. Both options are too costly for most African companies. Nevertheless, in developed

countries, the cloud might remain as a backup.

“We sort of build the technology with the mindset that the internet does not exist.”

- Armin Said -

A third opportunity is the ability of edge computing to preserve privacy. This ability is seen as an essential

advantage by industry experts, especially with regards to the General Data Protection Regulation (GDPR) in
25
Europe or the California Consumer Privacy Act (CCPA) in the State of California. Those specific regional

regulations make it necessary for companies to invest in security. As already mentioned, the main problem

people see in the cloud is that their data is stored somewhere far away. The storage might even be in a

different country where other privacy laws exist. Edge computing offers a solution as it filters and processes

the data closer to the end-user, which implies that the information does not have to be transferred to another

location via public networks. Edge computing provides flexibility to decide what data should leave the

device. Nevertheless, companies need to have a lot of education about the technology to ensure data security

in individual application cases. Especially, as the term itself is not intuitive. More research must, therefore,

be carried out and published by independent technology experts.

“Today the core challenge is one of education.”

- David Shackleton -

Finally, the ability to coordinate data on a local level is another opportunity for edge computing. One could

imagine the case of drones. Edge computing not only enables drones to understand their environment but

also enables them to cooperate, which could be helpful when searching for a suspect in a particular region.

In this case, people cannot rely on the cloud, as the reaction time needs to be reliable. In a less critical case,

however, drone data can be transferred to a central location. Overall, the question is where data processing

is needed and why. In general, there are different levels of coordination. Edge computing enables this

coordination locally.

“You may want to have drone exchanging information at a regional level.”

- Alberto Cresto -

4.2.2. Key Challenges of Edge Computing

After having mentioned three important arguments for edge computing, this section highlights four

challenges that must be addressed as the technology evolves. Overall, there are still many aspects that have

to be regarded more carefully. Nevertheless, those differ concerning their solvability.

26
One challenge, which all interviewees highlighted, is the diverse and sometimes even misguiding definition

of the term. Practitioners criticize the contradictory view on the term; thus, one could think of edge

computing as a local server, merely as sensors, or something that simply complements the traditional

network. Especially companies that are investing in the cloud tend to view everything in the cloud. However,

the market needs to be educated about the exact meaning of edge computing concerning specific applications.

This will get the benefits out of the technology. The critical question concerning use cases is what is critical

and what has to run locally. Besides, respondents highlight the lack of consolidation of communication

protocols, the lack of coherence and standardization, as a significant obstacle. Overall, this challenge will be

less of an issue as this is currently being solved.

“There's a lot of people talk about Edge, but it can


mean hundred and fifty thousand different things.”

- Jurgen Hofkens -

Before looking at the challenge with data management, the security chain is vital to consider in advance.

Considering an Industrial Internet of Things (IIoT) application, there is a need to connect all devices with

the right quality of service securely. In a digital fabric, this would imply orchestrating all the data from

different component providers, each of them having different types of hardware. However, the problem is

the heterogeneous nature of edge computing, providing little information about the interacting device. This

adds critical abstraction to the process. Thus, interviewees that offer an edge computing product or service

question whether it is possible to provide edge computing without having access to the customers´

confidential data. It seems impossible to integrate all components of multiple network providers securely.

Besides, the cost of integration alone would be huge. Other than that, there is a challenge in physical security

as well, as edge hardware is very close to people and objects. Hence, criminals could simply steal the object

and will get access to critical data. Overall, the security issue is much more challenging to solve than the

first, the misguiding definition.

“Edge by definition is heterogeneous. So […] you have very little control


or insight into what sort of devices you may be interacting with.”

- Paul Mundt -

27
As soon as there is sufficient security in place, the next major challenge with edge computing is data

management, especially when there is no managed infrastructure in the location where people want to deploy

it, for example, in emerging countries. There are a few companies that are trying to provide solutions to

specific edge issues on top of the cloud; however, mostly, they assume a managed infrastructure underneath.

As described earlier, the cloud comprises information from multiple computers, each having its own

intelligent management system. In order to configure and control this data, there needs to be a smart

management system in place. Edge computing technology can virtualize and manage directly on the edge.

However, it is challenging to be at all individual locations; for example, in the case of Circle K. Thus, there

also needs to be intelligent management in the local edge server, and being software-defined is crucial. It is

questionable if this requirement is given in every IoT use case. Therefore, this challenge might not be solved

in every case.

“Once all the security is done, doing the data management


and doing the remote management of the devices is a challenge.”

- Jurgen Hofkens –

“Being software defined is a critical technology enabler.”

- Brian Buggy -

Country specific regulation is an additional challenge of edge computing, especially concerning autonomous

cars. If a car aggregates data in one county, this might not comply with the law in another country.

Interviewees mentioned the GDPR, which has been introduced in 2018 and harmonizes data privacy laws

across Europe (Intersoft Consulting, 2020). Hence, data analysis and cross-border data flow between

member states and non-European countries are only possible to a limited extend. When a person travels to

another country with his car, certain features may need to be turned off as they may not comply with the

laws of the other country. Alternatively, the system would need to be turned off completely. For instance, if

companies want to perform analysis across different data sets. The firm needs a place where they can move

the query down to run it in the different locations and then aggregate the results of the query rather than

moving the data itself. Overall, it is evident that regulation here is hampering innovation and, in particular,

28
the development of edge computing. In this case, the possibility of having a cross-border standard seems

almost insoluble.

“As the car moves around, it creates all sorts of compliance


headaches for the data flows and for the back end.”

- Paul Mundt -

4.3. The Edge Computing Market

After an assessment of the origin and capabilities of edge computing in the previous sections, the following

abstract attempts to illustrate the edge computing market, its maturity, and growth, as well as its key players.

On this basis, a conclusion can be drawn about Entrepreneurial Ecosystems, which will be evaluated in more

detail in the discussion.

4.3.1. Market Growth and Maturity

When thinking about the ecosystem of edge computing, one should consider the maturity of the technology.

According to several interviewees, the time to invest in edge computing was five to seven years ago. At that

time, the term “edge computing” did not yet exist, and the industry expected that all computing would take

place in the cloud. Currently, companies, policymakers, or investors are becoming increasingly aware of the

possibilities of the technology. This attention of the public indicates that the market is growing and heading

towards maturity.

“Edge computing is now becoming mainstream.”

- Brian Buggy -

In theory, the development of products and services can be explained by the Product Life Cycle in Figure

11, which shows a clear pattern over time. After the implementation of a new product on the market, the first

competition is low, and the market size is small. This phase is typically the opportunity for small, agile, and

disruptive companies that detect a niche in the market. Over time, the product or service becomes

increasingly attractive to customers and peers. The market size keeps increasing until maturity. This

theoretical explanation seems clear and logical. The right time for investment and market entry is very crucial

for start-ups, but in practice, this is also the biggest challenge. Besides, interviewees emphasize that the

29
Technology Adoption Cycle does not uniformly take place throughout the market, which makes the decision

found a start-up even more difficult.

Figure 11: The Product Life Cycle

“There are some parts in the


industry that are early adopters
and some that are late adopters.”

- Brian Buggy -

“The major issue with most things


are what's called the technology
adoption life cycle.”

- Brian Buggy -
Source: Knowledge Grab (2020)

Figure 12: Global Edge Computing Market


“All of that technology, today is
USD 22,153.9 Million
Chinese or American. There is no
European technology. Zero.”

- Jurgen Hofkens -

“In Europe […] it looks like


looking in the US one year ago.”
2017 2018 2019 2020 2021 2022 2023 2024

- Alberto Cresto -
Source: Market Research Future (2019)

Based on these findings, it becomes evident that the attention of edge technology is increasing. This increase

in attention is also positively correlated with the growth of the market. The increase in the market size can

be seen in Figure 12, which illustrates the expected market growth from 2017 until 2024. Similarly, the

interviewees believe that the edge computing market will be huge and that it will grow. Another market

report by MarketsandMarkets (2019) states that North America, Asia Pacific, and Europe are among the

major geographical regions for edge computing. The main drivers behind this growth are new technologies,

like 5G, IoT, VR, AR, or AI, producing large amounts of data that need to be processed intelligently. In

contrast to this market report, interviewees question the existence of the European edge computing market,

claiming that the technology is all Chinese or American. Indeed, when thinking about technology hubs, one

would instead think about Tel Aviv in Israel; Shanghai, Bangalore, Hong Kong or Singapore in Asia; or

30
Silicon Valley in the US. In Europe, only Berlin or London would probably be the most attractive hubs for

technology start-ups. The stable technology environment in Asia or the US might lead to more substantial

growth and innovative potential compared to Europe.

Does the edge computing market already indicate the existence of hot spots? In general, competition is fierce

for market participants at the time when hot spots can be identified. Besides, hot spots indicate that a product

or service is already on the commercial side. As a result, the ability to invest resources in a specific product

or service is more limited for small businesses. When looking at the results from Crunchbase (Figure 13 &

14), it becomes clear that there is neither a large number of start-ups in Europe nor in the US. However, in

Europe, edge computing start-ups are mainly located in the United Kingdome, as well as in Germany. In the

US, the majority of the edge computing start-ups are located in California. These observations fit with the

assumption of the previous paragraph.

Figure 13: Edge Computing Start-Ups in Europe

United Kingdom: 9
Germany: 5
The Netherlands: 2
Switzerland: 2
Finland: 1
Spain: 1
Sweden: 1
Belgium: 1
Malta: 1

Sum: 23
Source: Crunchbase

Figure 14: Edge Computing Start-Ups in the United States

California: 27 New Jersey: 1


Texas: 3 Missouri: 1
Washington: 2 Florida: 1
Massachusetts: 2 Arizona: 1
Utah:1 Colorado: 1
Tennessee: 1 Oregon: 1
North Carolina: 1
Illinois: 1

Sum: 43

Source: Crunchbase

31
4.3.2. Key Players on the Market

Currently, the edge computing market consists of various players, from small niche vendors, or start-ups, to

existing cloud providers - such as Amazon and Microsoft - which have their edge computing capabilities

and are snowballing. Currently, there seems not to be one player that owns and offers one global edge device.

Nevertheless, a few companies are trying to dominate, like MobiledgeX from Telekom or Edge Gravity from

Ericsson. On the one hand, interviewees describe this tendency towards the “winner takes all” as a typical

characteristic of the technology industry. Such winners are, for example, companies like Microsoft with its

Windows offering, or Amazon, dominating in e-commerce or cloud computing. On the other hand, others

believe that with edge computing, there will be no winner. As already mentioned, one key challenge of edge

computing is the management of data, if there is no managed infrastructure underneath. In addition to that,

edge computing is heterogeneous by nature. Hence, finding one player, who can orchestrate all devices and

components, for example, in the case of IIoT, is almost impossible. Besides, one could think about a situation

where multiple players play different roles. Policymakers, start-ups, incumbents, or investors could

potentially partner up and work together to enhance and improve the connectivity infrastructure.

At the moment, there is some kind of consolidation taking place. The edge computing market attracts

established firms like Amazon, Microsoft, Cisco, or Huawei. Those have been mainly investing in the cloud

before and are located outside of Europe. In order to stay competitive, incumbents are now starting to buy

promising technology from start-ups to scale their business quickly. The main advantage of established cloud

providers, like Amazon, is that they can now offer hybrid solutions, including both edge and cloud

technology. Simultaneously they create more dependence on their offering. However, as already mentioned

before, people do not want to have this dependence. Besides, by offering a hybrid solution, incumbents do

not reinvent anything; they simply displace specific tasks. In addition to this argument, offering hybrid

solutions requires established cloud providers to scale down their existing business to small servers on the

edge. This might be difficult and costly.

“Consolidation is starting to occur.”

- Brian Buggy -

32
“All of this innovation is coming from start-ups.”

- Alberto Cresto -

4.4. The Funding Landscape of Edge Computing Ecosystems

The market overview in the previous part gave a feeling for the market maturity, its growth, and its key

players. The following part will delve deeper into the financing landscape and clearly distinguish the

financing opportunities in the USA from those in Europe.

When talking about investments, the type of investment should be clearly stated. On the one hand, start-ups

invest their time and know-how mainly on a specific topic. On the other hand, financial investors primarily

invest money into promising new ventures. Furthermore, there is also a difference in investment culture

among countries, which will be discussed between the US and Europe.

Looking at the location of companies that invest in edge computing start-ups, it can be seen that they are

located outside Europe. Respondents criticise the fact that there is simply no medium to large technology

firms in Europe. Hence, there is a lack of corporate investment and, simultaneously, the lack of an exit

environment.

“80, 85, maybe even 90 percent of the potential buyer are in the states.”

- Alberto Cresto -

In general, the demand for digital Europe is growing. Thus, the European Commission has to act and put

money in the industry, concerning the fierce competition from Asian countries, as well as from the US. At

the moment, the European Commission created an environment that is highly competitive for

telecommunication firms and decreases their margins. Therefore, there is no dominant player in Europe,

which could drive an edge computing ecosystem, at least not a mobile operator. On the one hand, the lack

of one dominant player means that the same service is created in every country with different parties. Hence,

it takes a lot longer before a stable environment is created. On the other hand, the end-user has a variety of

options to choose from, and more innovation is created in the market. In the United States, the situation is

entirely different. AT&T and Horizon dominate the telecommunications market, meaning that they could

33
potentially create and dominate the edge computing market. Both companies have the power to tell small

suppliers what to do, which makes the ecosystem faster, but also less innovative. The advantage for start-

ups in the US is that they can conquer the whole market when working together with AT&T or Horizon. In

Europe, this would be impossible, as start-ups cannot work in 27 countries and support 27 different

ecosystems at the same time.

“Today, we need to build a digital Europe.”

- Jurgen Hofkens -

“We need to build a critical infrastructure. And I think


edge computing in Europe is an opportunity.”

- Jurgen Hofkens -

Overall, the lack of European capital is being criticized. Interviewees similarly describe the challenge of

technology start-ups to get money in Europe. Furthermore, they stress that moving their business to

California would probably facilitate fund-raising of hundreds of millions of dollars and fail. In Europe, it is

difficult or even impossible to find 10 million dollars with the same kind of power. To illustrate this, one

could think about Pensando Systems; a US start-up founded a year and a half ago by the ex-CEO of Cisco.

Presently they have little technological strength and know-how. However, they have already raised $278

million (Bort, 2019). Thus, the chance that they will overtake their European peer’s technology-wise is high.

The reason is that they are now able to hire a thousand engineers.

“If you start and the Americans spot something that you're doing, they
can get more money on the problem faster than you can.”

- Brian Buggy -

Looking at the average financing of start-up companies that focus specifically on edge computing, a small

difference can be seen between North American and European companies (Appendix 1 & 2). The average

funding of American edge computing start-ups, founded after 2010, presently is $ 8.104.356,13, whereas

average European firms, founded after 2010, raised $ 5.799.912,43 on average. Those numbers underline

the observations of the interviewees; however, due to the small sample size, it is questionable if they

34
represent the real population mean. Crunchbase identified 43 start-ups in the United States (24 have data on

the total funding amount), and 23 in Europe (seven firms have data on the total funding amount).

Two further aspects should be taken into account to explain why it is more challenging to raise funds in

Europe. On the one hand, there is a difference in the investment culture. American investors have a strong

focus on the vision, the mission, and the team behind the start-up. Conversely, European investors look at

the revenues and profits. On the other hand, the willingness of end-users to try out new technologies in

Europe is low, although there are some advanced manufacturing companies.

35
5. Discussion

In the following part, I will first sum up the key implications of my thesis. Based on my analysis, I was able

to indicate essential factors influencing the edge computing ecosystem. Thus, I will link specific elements

of entrepreneurial ecosystems with the findings from my qualitative analysis. Afterwards, I will give some

practical recommendations for policymakers and practitioners. In the end, I highlight limitations to my

research and suggest how edge computing can be further examined.

5.1. Summary of the Findings

I have learned that an entrepreneurial ecosystem is “a set of interdependent actors and factors coordinated in

such a way that they enable productive entrepreneurship” (Stam, 2015, p. 1765). Hence, I believe that

knowledge of the key actors and factors within a given region is crucial to identify and assess an ecosystem.

The key actors that I identified within the entrepreneurial ecosystem of edge computing are financial

investors, policymakers, and edge computing vendors. Edge computing vendors can be divided into two

distinct groups. On the one hand, there are edge computing start-ups that offer specific niche solutions and

focus on one to three applications. On the other hand, there are large corporates, like Amazon or Microsoft,

that offer hybrid solutions. The key factors within the entrepreneurial ecosystem of edge computing are

accessible markets, support systems, regulatory framework, finance, and culture. Those factors enable

productive entrepreneurship and differ between locations, for example, between Europe and the US. When

looking at the accessible market, it is difficult to find European customers that are willing to be the first to

incorporate new technologies in general. In the US, this is not the case. Thus, the European market is less

accessible for edge computing solutions. Likewise, as an edge computing start-up in Europe, it is less

probable to find a reliable support system. Most people expect promising technological ideas to come from

California, as there is also a high density of edge computing entrepreneurs. In the US, most of the edge

computing start-ups (27 out of 43 edge computing start-ups in total) are located in California. In Europe, the

majority of all edge computing start-ups are based the UK and Germany. Both countries together comprise

only 14 edge computing start-ups. However, compared to California, this figure seems almost irrelevant.
36
Overall, the regulatory framework is decisive for competition within a region. At the moment, the European

Commission creates fierce competition among companies. On the one hand, competition creates value; on

the other hand, it decreases margins. Besides, in the US, it is more likely to have one or two dominant edge

computing vendors in the near future that could potentially drive an ecosystem. Furthermore, there is a lack

of capital in Europe. One reason why investors in the US invest significantly more capital than their European

counterparts is their positive attitude towards risk and failure. European investors are focussed on profits and

are less willing to take a risk or tolerate failure. Thus, one can also observe a difference in the investment

culture. The European attitude towards risk could be due to the fierce competition from the USA, and more

generally, to the suboptimal ecosystem environment. The lack of cultural support might also be caused by

the lack of European success stories or role models.

I conclude that there is no digital Europe in place at the moment. There is a realistic probability that, once

again, America or Asia outperform Europe - for example, considering the cloud computing market with US-

based giants, like Amazon, Microsoft, and Google (Ranger, 2019). However, edge computing might offer a

possibility to change this situation. This particular technology is crucial in many new applications, be it in

autonomous driving, retail, or drones. Furthermore, edge computing can shift the dependence on large cloud

providers away and enable more intelligence at the edge. In order to take a leading role and build a unique

selling proposition in Europe, I will now make concrete recommendations to European policymakers and

practitioners.

Figure 15: Key factors within the entrepreneurial ecosystem of edge computing (by region)

Source: Own Creation

37
5.2. Implications for Policymakers and Practitioners

One should keep in mind that edge computing use cases have diverse requirements. Besides, the

implementation of edge computing depends on the respective market and its circumstances. I decided to

focus my implications on policymakers and practitioners, as they represent both key actors within the

entrepreneurial ecosystem of edge computing. Furthermore, to narrow down my recommendations, I will

focus on those implications on Europe.

5.2.1. Implications for Policymakers

In my opinion, politicians have the responsibility to spot promising new technology and support start-ups.

When talking about the potential of an edge computing ecosystem in Europe, it is highly relevant to consider

all niche players instead of looking exclusively for high-growth start-ups. The reason why I think

policymakers should consider all players is that edge computing start-ups mostly have their particular

application area. Nevertheless, the government should support especially those ventures that, for instance,

have valuable human resources but lack other things to develop their innovation further. Often, they require

financial capital, or their operations are limited by specific regulations, like the GDPR.

The European Commission will have to decide if it adopts supporting policies for entrepreneurial

ecosystems. There are multiple possibilities to build critical infrastructure, for example, subsidising edge

computing start-ups, supporting technology-oriented studies in general, or hosting start-up competitions with

the focus on edge computing. In particular, by setting a standard to facilitate cross-border data analyses, the

government can advance the technology in autonomous cars. If Germany, for example, wants to keep its

leading position in the car industry, the government has the responsibility to advance edge computing

technology. I recommend collaboration between car manufacturers, the government, and specialized edge

computing start-ups. Together, those players can build a leading position. Besides, the government should

collaborate with other countries to facilitate data analysis across borders.

Governments also have the power to decide whether they create intense competition within the ecosystem

or advance the emergence of a dominant payer within the market. One possible scenario for the future of

38
edge computing would be the emergence of one powerful entity that acquires promising start-ups for each

application area. As discussed in part 4.4., this would be possible in the US. In the US, the competitive

environment is not as intense as in Europe. Thus, there are many dominant technology companies. With

regards to edge computing, either the two major telecommunication companies, AT&T and Horizon, win

the game or large cloud providers, such as Amazon and Microsoft. Another scenario would be an extensive

network of edge computing companies sharing their experience and expertise on one platform while running

their businesses independently. This platform could be provided by the government, facilitating the

interaction and exchange between the public sector and entrepreneurs. As Stam (2015) and Teece (1992)

emphasised, non-market interactions are as crucial for the realisation of innovation as market interactions.

As opposed to the first scenario, I think that the second scenario would be more likely to happen in Europe.

As many new edge computing use cases are not being deployed at the moment, a platform, as such, could

create motivation to try out new things and jointly work together on projects, proposed by the government.

At the same time, this would enhance the support system within Europe. This is particularly important as

European culture tends to be risk-averse.

5.2.2. Implications for Practitioners

Practitioners can be divided into financial investors and edge computing start-ups. Both parties need to be

educated about the technology. Education about the specific capabilities of edge computing, as well as its

diverse definition, ensures targeted investments. Either financial investment in unique start-up ideas or

investment in terms of putting effort into the development of one specific application. A network, driven by

the government, could be advantageous to share knowledge and experience.

What are the current edge computing use cases, and where will the technology be applied in the future?

There is no doubt that the possibilities to utilize edge computing are enormous. Edge computing could

theoretically be used wherever there is connectivity with multiple users. Furthermore, edge computing is

used in applications where users have different privacy preferences, and different requirements. However,

the relevant applications will be in areas where latency and privacy play an essential role.

39
Nonetheless, not every use case is being applied, even if many people talk about them. It seems as if

everybody knows that they need edge computing, but hardly anybody wants to be the first to deploy. At least

in some use cases. This can be attributed to the risk aversion of Europeans (Cipollini et al., 2018). As an

edge computing start-up, people should be willing to take a risk and accept failure. Also, being the first

mover creates many opportunities to build valuable knowledge about a particular application and become

the leader of a niche. On the other hand, those who are the first to implement something have no standard or

reference point, which can be difficult. Being the second mover could be an advantage because they can

learn from first movers and can easily follow in the footsteps. However, start-ups then have to compete with

financially stable incumbents. Competing with incumbents on the same market is almost impossible for start-

ups.

Currently, there is much hype around edge computing, especially concerning IoT. Many people believe that

humankind will soon live in a fully autonomous, smart, and connected world. Many research reports mention

the early 2020s. However, full connection and automatization will probably happen in ten to fifteen years.

Hence, the timeframe illustrated in many research reports is unrealistic and overambitious. The main

problem with IoT is the assumption that edge computing supports it and that it is mature and ready for

deployment. However, in reality, the two are symbiotic and even work off one another. IoT develops new

requirements, and edge computing has to evolve to support those requirements. IoT requires edge computing

everywhere, not just in physical locations. Besides, the business case will not always stack up, for example,

considering the value of the data source. It might be possible from a technological perspective to deploy, but

would it make sense to put edge computing into light bulbs? This value limit is also stopping the IoT.

Furthermore, the physical size of the computer needed for edge computing continually decreases. However,

the management software can only shrink to a certain extent, as it needs to deliver the software, ensure

security, manage several processes, or gather data at the same time. IoT technology wants edge computing

to go everywhere, but what happens if the computer is not able to shrink small enough to support the

application? This is a problem of scaling down. In my opinion, it is difficult to solve this from a current point

40
of view. If edge computing start-ups want to go into IoT, I would recommend collaborations with incumbents

to secure financial support and develop a solution.

Edge computing start-ups should keep in mind that many things will still rely on the cloud in the future. One

example would be in IIoT applications. Being only in the cloud market puts a company into the mainstream

part of the market with a lot of competition. Whereas being solely in the edge computing business puts the

firm into a niche. Nowadays, there is some kind of hybrid model needed. As already discussed, this can be

offered by large corporates. Therefore, incumbents take advantage of specific edge computing know-how

from small ventures. Concerning the disruptor’s dilemma, start-ups should decide to either work together

with incumbents or develop a leading position in a specific niche.

In general, individual firms should keep in mind that their ability to create value depends on structural and

strategic factors. As mentioned in part 5.1. technology start-ups in Europe face a different situation than the

ones in the US. In particular, the different policies hinder cross-border data analysis within the application

of autonomous cars. Besides, edge computing start-ups have to bear in mind that they have limited resources,

like human capital or financing, compared to incumbents. Thus, in my opinion, it does not make sense for

edge computing start-ups to work on more than three application areas at once. Moreover, start-ups should

try to become a leader in one or two specific niches and scale their business quickly from there.

Cooperation with several parties, especially in a situation such as the current Covid-19 crisis, can raise

awareness of edge computing technology. Presently, there is a promising opportunity for start-ups,

incumbents, and other organizations to work together on a specific use case. Within a German project

between SAP, Telecom, as well as several start-ups, the parties jointly work together on a decentral tracing-

app for Covid-19 suspects (Scheuer and Klöckner, 2020). The idea behind this application was to preserve

the privacy of app users (Turzer, 2020), as personal data will be stored decentral on the user’s device.

Bluetooth signals can trace back users that have been positively tested without sharing personal details. For

German edge computing start-ups, this situation is an opportunity to become more independent of US

technologies in the future and to prove that the idea of decentral data processing has enormous potential.

41
To become successful in the edge computing market, start-ups should be aware of their location, as well as

about the awareness within the market. There might be a broader accessible market in the UK and Germany

than in the rest of Europe. It could also be interesting to analyse peers within related markets to spot a niche.

Therefore, I suggest examining figure 13 as well as Appendix 1. Besides the possibility within the European

market, I hold the opinion that emerging countries are new markets to apply edge computing technology, for

example, Africa. Reasons for that are the missing infrastructure that requires local data management, a lot

of innovation potential, and the urgent need for technological advancement. European edge computing start-

ups could test applications in such an extreme environment and see if the technology is working without any

infrastructure underneath. Another reason to invest in emerging markets is that they are a niche, and there is

currently no significant competition from large cloud vendors.

5.3. Limitations of the Thesis and Opportunities for Further Research

The most critical part of this thesis involved evaluating the available data and finding a logical structure.

Drawing a conclusion from the Crunchbase output table and linking it to the statements of interviewees

presented a challenge. Especially as the data on the total funding amount was not available for every

company. Besides, there are not many edge computing start-ups in general.

Due to the limited amount of time, it was also not possible to look up every company individually. Thus, I

decided to solely contact and interview European edge computing entrepreneurs and market experts.

Therefore, all statements about the difference between American and European investments are made from

a European perspective.

As mentioned before, the term edge computing is not intuitive. Merely listening to large and vocal cloud

providers can be misleading, especially since I do not have a deep technological understanding. For instance,

cloud provides cannot imagine a world without the cloud. For them, edge computing is an attractive add-on.

Hence, they create a hybrid solution for their customers. Talking to diverse entrepreneurs and market experts

helped to understand the topic in more debt and recognize the diversity of edge computing. Nevertheless, it

42
took a considerable amount of effort to find relevant and matching comments, as each expert I interviewed

focused his business on a different application.

Future research could set a more in-depth focus on a specific use case and discuss complementary

technology, like AI or 5G. For example, does edge computing depend on the advancement of AI or could

the technology work without it?

Besides, future research could focus on the capital investor perspective. In my dissertation I gave a brief

overview of this, but I only interviewed one financial investor. Moreover, my Crunchbase data was not

complete concerning information on the total funding amount.

43
References

Acs, Z. J., & Audretsch, D. B. (1988). Innovation in large and small firms: an empirical analysis. American
Economic Review. Pp. 678–690.
Acs, Z. J., Astebro, T., Audretsch, D., & Robinson, D. T. (2016). Public policy to promote entrepreneurship:
a call to arms. Small Business Economics. Pp. 35-52.
Acs, Z. J., Stam, E., Audretsch, D. B. & O’Connor, A. (2017). The lineages of the entrepreneurial ecosystem
approach. Small Business Economics. Springer.
Adner, R., & Kapoor, R. (2010). Value creation in innovation ecosystems: how the structure of technological
interdependence affects firm performance in new technology generations. Strategic Management
Journal. Pp. 306–333.
Adner,R., Oxley, J. E. & Silverman, B. S. (2013). Collaboration and competition in business ecosystems.
Advances in Strategic Management. Vol. 30.
Alvedalen, J., & Boschma, R. (2017). A critical review of entrepreneurial ecosystems research: towards a
future research agenda. European Planning Studies. Pp. 887-903.
Amin, A., & Thrift, N. (1994). Globalization, institutions, and regional development in Europe. Oxford:
Oxford University Press.
Ansari, S., Kumaraswamy, A. & Garud, R. (2015). The disruptor's dilemma: TiVo and the U.S. television
ecosystem. Strategic Management Journal.
Anselin, L., Varga, A., & Acs, Z. J. (1997). Local geographic spillovers between university research and
high technology innovations. Journal of Urban Economics. Pp. 422–448.
Autio, E., & Thomas, L. (2014). Innovation ecosystems. The Oxford Handbook of Innovation Management.
Oxford: Oxford Universitys Press. Pp. 204-288.
Baumol, W. J. (1990). Entrepreneurship: Productive, unproductive, and destructive, Journal of Political
Economy. Pp. 893–921.
Becattini, G. (1990). The Marshallian industrial district as a socioeconomic notion. In F. Pyke, G. Becattini,
&W. Sengenberger (Eds.), Industrial districts and inter-firm co-operation in Italy. Geneva:
International Institute for Labour Studies.
Bengtsson, M. & Kock, S. (2000). “Coopetition” in business networks—to cooperate and compete
simultaneously. Industrial Marketing Management. Pp. 411–426.
Bort, J. (2019). John Chambers and a star team of ex-Cisco engineers have finally launched Pensando
Systems, a startup with $278 million in funding, to take on Amazon — and Cisco. Retrieved on
05.05.2020 from: https://www.businessinsider.de/international/john-chambers-pensando-systems-
cisco-stars-amazon-2019-10/?r=US&IR=T
Bower, J. L. & Christensen, C. M. (1995). Disruptive Technologies: Catching the Wave. Harvard Business
Review. Pp. 43-53.

V
Brandenburger, A. & Nalebuff, B. (1996). Coopetition: A Revolution Mindset That Combines Competition
and Cooperation: The Game Theory Strategy That’s Changing the Game of Business. Currency
Doubleday: New York.
Bruns, K., Bosma, M. & Sanders, M. (2017). Searching for the existence of entrepreneurial ecosystems.
Small Business Economics. Springer. Pp. 31-54.
Caillaud, B., & Jullian, B. (2003). Chick and egg: competing matchmakers. RAND Journal of Economics.
Pp. 309-328.
Charitou, C. & Markides, C. (2003). Responses to Disruptive Strategic Innovation. Sloan Management
Review.
Chemitiganti, V. (2019). Edge computing challenges and opportunities. Retrieved on 01.04.2020 from:
https://internetofthingsagenda.techtarget.com/blog/IoT-Agenda/Edge-computing-challenges-and-
opportunities
Christensen, C. M. & Bower, J. L. (1996). Customer power, strategic investment, and the failure of leading
Firms. Strategic management journal. Pp. 197-218.
Christensen, C. M., Anthony, S. D., & Roth, E. A. (2004). Seeing What´s Next. Harvard Business School
Press.
Christensen, C. M., Raynor, M., & McDonald, R. (2015). What is disruptive innovation?. Harvard Business
Review.
Cipollini, A., Cascio, I. L., Muzzioli, S. (2018). Risk aversion connectedness in five European countries.
Economic Modelling. Elsevier. Vol. 71. Pp. 68-79.
Coase, R. (1937). The nature of the firm. Economica, new series. Pp. 386–405.
Cooke, P., Gomez Uranga, M., & Etxebarria, G. (1997). Regional innovation systems: institutional and
organizational dimensions. Research Policy. Pp. 475-491.
Crunchbase (2020). Who we are – Company. Retrieved on 23.04.2020 from:
https://about.crunchbase.com/about-us/
Davis, A., Parikh, J. and Weihl, W. E. (2004). EdgeComputing: Extending Enterprise Applications to the
Edge of the Internet. Akamai Technologies. New York, USA.
Davis, S., Haltiwanger, J., & Schuh, D. (1996). Job creating and job destruction. Cambridge: The MIT Press.
Dilley, J., Maggs, B., Parikh, J., Prokop, H., Sitaraman, R. & Weihl, B. (2002). Globally Distributes Content
Delivery. Akamai Technologies. IEEE Internet Computing.
Dini, P., Iqani, M., & Mansell, R. (2011). The (im) possibility of interdisciplinary lessons from constructing
a theoretical framework for digital ecosystems. Culture, theory and critique. Pp. 3-27.
Evans, D. S., & Schmalensee, R. (2016). Matchmakers: the new economics of multisided platforms. Boston:
Harvard Business Review Press.
Evans, D. S., Hagiu, A. & Schmalensee, R. (2006). Invisible Engines: How Software Platforms Drive
Innovation and Transform Industries.MIT Press: Cambridge, MA.
Gilbert, C. & Bower, J. L. (2002). Disruptive Change: When trying harder is part of the problem. Harvard
Business Review.

VI
Gnyawali, D. & Park, B. (2011). Coopetition between giants: collaboration with competitors for
technological innovation. Research Policy. Pp. 650-663.
Heineke, K., Kampshoff, P., Mkrtchyan, A., Shao, E. (2017). Self-driving car technology: When will the
robots hit the road?. McKinsey & Company.
Iansiti, M. & Levien, R. (2004). The Keystone Advantage: What the New Dynamics of Business Ecosystems
Mean for Strategy, Innovation, and Sustainability. Harvard Business School Press: Cambridge, MA.
IBM (2020). What is Edge Computing?. IBM. Retrieved on 08.03.2020 from:
https://www.ibm.com/cloud/what-is-edge-computing.
Intersoft Consulting (2020). General Data Protection Regulation. Retrieved on 21.04.2020 from:
https://gdpr-info.eu/
Isenberg, D. J. (2010). How to Start an Entrepreneurial Revolution, Harvard Business Review. June. Pp. 41-
51.
Isenberg, D. J. (2011) Introducing the entrepreneurship ecosystem: Four defining characteristics, Forbes.
Available at http://www.forbes.com/sites/danisenberg/2011/05/25/introducing-the-
entrepreneurship-ecosystemfour- defining-characteristics/
Jacobides, M. G., Knudsen, T. & Augier, M. (2006). Benefiting from innovation: value creation, value
appropriation and the role of industry architectures. Research Policy. 35(8). Pp. 1200–1221.
Ketchen, D. J., Snow, C. C. & Hoover, V. L. (2004). Research on competitive dynamics: recent
accomplishments and future challenges. Journal of Management 30. Pp. 779–804.
Knowledge Grab (2020). Product life cycle. Retrieved on 23.04.2020 from:
http://knowledgegrab.com/glossary/product-life-cycle-2/
Li, W., Badr, Y. & Biennier, F. (2012). Digital ecosystems: challenges and prospects. In Proceedings of the
International Conference on Management of Emergent Digital EcoSystems. Pp.117–122. ACM.
Market Research Future (2019). Edge Computing Market Research Report - Global Forecast till 2024.
Retrieved on 19.04 2020 from: https://www.marketresearchfuture.com/reports/edge-computing-
market-3239
MarketsandMarkets (2019). Edge Computing Market by Component (Hardware, Platform, and Services),
Application (Smart Cities, IIoT, Content Delivery, Remote Monitoring, AR and VR), Organization
Size (SMEs and Large Enterprises), Vertical, and Region - Global Forecast to 2024. Retrieved on
08.03.2020 from: https://www.marketsandmarkets.com/Market-Reports/edge-computing-market-
133384090.html
Marshall, A. (1920). Principles of economics. London: MacMillan and co..
Mason, C. & Brown, R. (2014). Entrepreneurial ecosystems and growth oriented entrepreneurship.
Background paper prepared for the workshop organised by the OECD LEED Programme and the
Dutch Ministry of Economic Affairs on Entrepreneurial Ecosystems and Growth Oriented
Entrepreneurship, The Hague, Netherlands.
Matthews, K. (2018). 6 Challenges to be overcome for Edge Computing to work. Irish Tech Times. Retrieved
on 02.04.2020 from: https://irishtechnews.ie/6-challenges-to-be-overcome-for-edge-computing-to-
work

VII
Moazed, A., Johnson, N,. (2016). Why Clayton Christensen is wrong about Uber and disruptive innovation.
TechCruch. Retrieved on 17.02.2020 from: https://techcrunch.com/2016/02/27/why-clayton-
christensen-iswrong-about-uber-and-disruptive-innovation/
Nye, J.S. (2004). Soft Power: The Means to Success in World Politics. Public Affairs: New York.
O’Reilly, C.A. & Tushman, M.L. (2011). Organizational ambidexterity in action: how managers explore and
exploit. California Management Review. 53(4). Pp. 5–22.
Panetta, K. (2017). “Gartner Top 10 Strategic Technology Trends for 2018”. Gartner. Retrieved on
08.03.2020 from: https://www.gartner.com/smarterwithgartner/gartner-top-10-strategic-technology-
trends-for-2018/
Parker, G. G. & van Alstyne, M. W. (2005). Two-sided network effects: a theory of information product
design. Management Science. 51(10). Pp. 1494–1504.
Pitelis, C. (2012). Clusters, entrepreneurial ecosystem co-creation, and appropriability: a conceptual
framework. Industrial and Corporate Change. 21(6). Pp. 1359–1388.
Porter, M. (1990). The competitive advantage of nations. New York: Free Press.
Porter, M. E. (1998). On competition. Boston: Harvard Business School Press.
Ranger, S. (2019). AWS, Microsoft or Google: Which cloud computing giant is growing the fastest?. ZDNet.
Retrieved on 16.05.2020 from: https://www.zdnet.com/article/aws-microsoft-or-google-which-
cloud-computing-giant-is-growing-the-fastest/
Reynolds, B. (2019). What is Edge Computing?. Baytech Consulting. Retrieved on 08.03.2019 from:
https://www.baytechconsulting.com/the-benefits-and-potential-of-edge-computing-baytech-
consulting/
Richter, C., Kraus, S., & Syrjä, P. (2015). The share economy as a precursor for digital entrepreneurship
business models. International Journal of Entrepreneurship and Small Business. 25(1). Pp. 18–35.
Rochet, J. C., & Tirole, J. (2003). Platform competition in two sided markets. Journal of the European
Economic Association. Pp. 990–1029.
Salesforce (2020). Where is my Salesforce instance located? Retrieved on 03.05.2020 from:
https://help.salesforce.com/articleView?id=000314281&language=en_US&type=1&mode=1
Scheuer, S. & Klöckner, J. (2020). Start-ups helfen SAP und Telekom bei der Entwicklung der Corona-App.
Retrieved on 07.07.2020 from: https://www.handelsblatt.com/technik/forschung-
innovation/kontaktverfolgung-start-ups-helfen-sap-und-telekom-bei-der-entwicklung-der-corona-
app/25806712.html
Schumpeter, J. A. (1934) The Theory of Economic Development. Cambridge, MA: Harvard University
Press.
Stam, E. (2015). Entrepreneurial Ecosystems and Regional Policy: A Sympathetic Critique. European
Planning Studies. Vol. 23. No. 9. Pp. 1759–1769
Stam, E., & Spigel, B. (2017). Entrepreneurial ecosystems. In R. Blackburn, D. De Clercq, J. Heinonen, &
Z. Wang (Eds.), The SAGE Handbook of Small Business and Entrepreneurship. London: SAGE.
forthcoming.

VIII
Stam, E., Bosma, N., Van Witteloostuijn, A., de Jong, J., Bogaert, S., Edwards, N., & Jaspers, F. (2012)
Ambitious Entrepreneurship. A Review of the Academic Literature and New Directions for Public
Policy (Den Haag: Adviesraad voor Wetenschap en Technologie-beleid (AWT)).
Stubbart, C. I. & Knight, M. B. (2006). The case of the disappearing firms: empirical evidence and
implications. Journal of Organizational Behavior. 27(1). Pp. 79–100.
Sussan, F. & Acs Z. J. (2017). The digital entrepreneurial ecosystem. Springer. Pp. 55-73
Tansley, A. J. (1935). The use and abuse of vegetational concepts and terms. Ecology. Pp. 284-307.
Teece, D. (1986). Profiting from technological innovation: implications for integration, collaboration,
licensing and public policy. Research Policy. Pp. 285–305.
Teece, D. (1992) Competition, cooperation, and innovation: Organizational arrangements for regimes of
rapid technological progress, Journal of Economic Behavior and Organization. Pp. 1-25.
Terjesen, S., Acs, Z. J., Audretsch, D. B., Hechavarria, D., Stam, E., & White, R. (2017). Entrepreneurial
ecosystems: the search for performance. University of Tampa.
Tilson, D., Lyytinen, K., & Sørensen, C. (2010). Research commentary-digital infrastructures: the missing
IS research agenda. Information Systems Research. Pp. 748–759.
Turzer, C. (2020). Bundesregierung favorisiert jetzt doch dezentrale Corona App. Retrieved on 07.05.2020
from: https://www.gruenderszene.de/technologie/bundesregierung-favorisiert-jetzt-doch-dezentrale-
corona-app
Uzunca, B., Sharapov, S., & Tee, R. (2016). Competition and cooperation in ecosystems: how industry
evolution and governance inseparability shape value capture over time. Working paper.
Van der Meulen, R. (2018). “What Edge Computing Means for Infrastructure and Operations Leaders”.
Gartner Research. Retrieved on 07.03.2020 from:
https://www.gartner.com/smarterwithgartner/what-edge-computingmeans-for-infrastructure-and-
operations-leaders/
Williamson, P., & De Meyer, A. (2012). Ecosystem advantage. California Management Review. Pp. 24–46.
World Economic Forum (2013) Entrepreneurial Ecosystems around the Globe and Company Growth
Dynamics (Davos: World Economic Forum).
Zahra, S. A., & Nambisan, S. (2011). Entrepreneurship in global innovation ecosystems. AMS Review. 1(1).
P. 4.

IX
Appendix

Appendix 1: Edge Computing Firms in Europe

Organization Name Location Industries Total Funding Amount Founded Date


Developer Platform, Developer
Tools, Internet of Things, SaaS,
Seashell Berlin, Germany Software 10.12.2019

Information Technology,
layline.io Hamburg, Germany Software 23.08.2019

Rain/ Lempea Edge Helsinki, Finland 01.04.2019


Artificial Intelligence, Big Data,
Almere, Flevoland, The Information Technology,
FactoryHub Netherlands Internet 01.04.2019

Cloud Computing, Developer


CloudBackend Ostergotlands Lan, Sweden APIs, IaaS, PaaS, SaaS, Software $ 820.000,00 19.03.2018

Artificial Intelligence, Enterprise


Software, Industrial, Industrial
Automation, Internet of Things,
London, England, United Machine Learning, Predictive
Ascalia Kingdom Analytics, SaaS, Smart Cities $ 196.421,00 01.03.2018

London, England, United Finance, Financial Services,


Luminous Ventures Kingdom Venture Capital 01.01.2018
Artificial Intelligence,
Information Technology,
Smartia Bristol, United Kingdom Software 01.01.2018
Data Center, Infrastructure,
Amsterdam, Noord-Holland, Internet of Things, Location
EdgeInfra The Netherlands
Marlow, Buckinghamshire, Based Services, Mobile 01.01.2018
edgeNEXUS United Kingdom Network Security 05.10.2017
Health Diagnostics, Medical,
Biospectal Lausanne, Vaud, Switzerland Medical Device, Wearables 06.07.2017
Artificial Intelligence, Cosmetic
Surgery, Machine Learning,
Medical, Software, Virtual
Arbrea Labs Zurich, Switzerland Reality 01.01.2017
London, England, United
EdgeX Foundry Kingdom Information Technology 01.01.2017
Artificial Intelligence,
Information Technology,
DiscoveryIoT Birkirkara, NA - Malta, Malta Software, Telecommunications 01.01.2016

CleanTech, Cloud Infrastructure,


Data Center Automation, Energy
Submer NextGen Efficiency, Hardware, IT
Datacenters Barcelona, Spain Infrastructure, Robotics
Augmented Reality, Industrial, $ 7.907.048,00 01.10.2015
Internet of Things, Software,
Holo-Light Ismaning, Bayern, Germany Virtual Reality, Wearables $ 4.746.851,00 01.04.2015
Lochristi, Oost-Vlaanderen, Cloud Computing, Information
GIG Technology Belgium Technology, Software 10.01.2015
Computer, CRM, Information
Virtuosys (Veea Inc.) Bath, United Kingdom Technology, Software 01.01.2015

X
Cloud Computing, Hardware,
London, England, United Internet of Things, Mobile,
WICASTR Kingdom Software, Wireless $ 650.000,00 01.01.2014

ENIT Energy IT Systems Freiburg, Baden-


GmbH Wurttemberg, Germany Energy, Renewable Energy 01.01.2014

London, England, United


Permutive Kingdom Publishing, SaaS, Software $ 11.529.067,00 01.01.2013

Core Network Dynamics Berlin, Germany Software 01.01.2013

Information Technology, Retail


Technology, Small and Medium
Zynstra Bath, United Kingdom Businesses, Software $ 14.750.000,00 30.11.2011

Appendix 2: Edge Computing Firms in the United States

Organization Name Location Industries Total Funding Amount Founded Date

Ciphense Riverside, California $ 75.000,00 20.02.2019

Artificial Intelligence, Hardware,


Industrial Automation,
Information Technology,
InsurTech, Logistics, SaaS,
Software, Supply Chain
KoiReader Technologies Dallas, Texas Management, Transportation $ 510.000,00 01.01.2019

Artificial Intelligence, Internet,


Celona Cupertino, California Mobile, Wireless $ 10.000.000,00 01.01.2019

Cloudbrink Inc. Sunnyvale, California Autonomous Vehicles, Cloud 01.01.2019


Management, Data Center
Automation, Developer Platform,
Internet of Things, Predictive
EDJX San Jose, California Analytics $ 2.000.000,00 01.09.2018

E-Commerce, Information
Difuon Lehi, Utah Technology,
Artificial Internet
Intelligence, 25.06.2018
Automotive, Information
Technology, Machine Learning,
Wavelength Global Dublin, California Software $ 100.000,00 01.01.2018

Cloud Computing, Cyber Security,


Cachengo Huntingdon, Tennessee File Sharing 01.11.2017

Artificial Intelligence,
Edgeworx San Jose, California Infrastructure, Internet of Things 27.09.2017

Industrial Automation, Machine


Compressor Controller Pittsboro, North Carolina Learning, Real Time 01.09.2017

Information Technology,
vitalpointz San Ramon, California Internet, Internet of Things 01.08.2017
Computer Vision, Hardware,
Internet of Things, Retail
SmaSS technologies Addison, Texas Technology $ 3.420.000,00 23.07.2017

XI
Cloud Computing, Data
Visualization, Education,
Information Technology,
ARICA IoT San Francisco, California Internet of Things, Software 01.03.2017
Information Services,
Information Technology,
Tignis Seattle, Washington Software $ 7.455.922,00 01.01.2017

Artificial Intelligence, Computer,


AtomBeam Technologies Moraga, California Machine Learning, Software $ 1.500.000,00 01.01.2017
Human Resources, Information
MobiledgeX San Francisco, California Technology, Internet 01.01.2017

Enterprise Software, Information


Logistyx Technologies Rolling Meadows, Illinois Technology, Software 01.01.2017
DinoPlusAI Fremont, California Artificial Intelligence 01.01.2017

TensorIoT Irvine, California Artificial Intelligence, Software 01.01.2017

Agriculture, Artificial Intelligence,


Drones, Internet of Things,
American Robotics Marlborough, Massachusetts Robotics $ 7.100.000,00 01.10.2016

Cloud Computing, Drones,


Internet of Things, SaaS,
Everduit Technology Sunnyvale, California Software 28.03.2016
Artificial Intelligence, Computer
Vision, Retail Technology,
AiFi Santa Clara, California Software $ 15.000.000,00 01.01.2016

quadric.io Burlingame, California Computer, Hardware $ 17.250.000,00 01.01.2016

Artificial Intelligence, Machine


Atlazo La Jolla, California Learning, Semiconductor $ 3.400.000,00 01.01.2016

Cloud Computing, Network


Spectronn Holmdel, New Jersey Security, Wireless $ 650.000,00 01.01.2016

Computer, Information
Nubix San Francisco, California Technology, Robotics, Software 01.01.2016

Advertising Platforms, Data


xSeer Santa Clara, California Integration, Video 01.01.2016

Association, Information
OpenFog Consortium Fremont, California Technology 15.09.2015
Automotive, B2B, Big Data,
Internet of Things, Machine
Learning, Predictive Analytics,
CarForce San Francisco, California SaaS, Software $ 2.950.000,00 21.04.2015

Flicq Inc. San Jose, California Internet of Things 01.04.2015


3D Technology, Augmented
Reality, Digital Entertainment,
Digital Media, Software, Video
Games, Virtual Reality, Virtual
DoubleMe Sunnyvale, California World $ 9.195.000,00 01.10.2014
Electrical Distribution, Industrial
Automation, Internet of Things,
Atomation Saint Louis, Missouri Logistics $ 10.206.000,00 09.06.2014

Analytics, Information
Olea Edge Analytics Austin, Texas Technology, Wireless $ 15.630.781,00 01.04.2014
XII
Analytics, Information
Olea Edge Analytics Austin, Texas Technology, Wireless $ 15.630.781,00 01.04.2014

Cloud Infrastructure, Enterprise


Software, Internet of Things, IT
Pixeom Santa Clara, California Infrastructure, Private Cloud $ 15.000.000,00 01.01.2014

Animusoft Miami, Florida Analytics, Enterprise Software $ 1.600.000,00 01.01.2014


Artificial Intelligence,
Information Technology,
Swarm Technology Mesa, Arizona Robotics 01.01.2014

Mobodexter Redmond, Washington Internet of Things $ 580.000,00 22.11.2013


Communications Infrastructure,
Data Center, Flash Storage, IaaS,
Aparna Systems, Inc. Fremont, California Software 01.01.2013
Big Data, Cloud Storage,
Computer, Data Storage, Flash
NGD Systems Irvine, California Storage $ 42.656.845,00 01.06.2013

Advertising, B2B, Digital Media,


SDxCentral Denver, Colorado Lead Generation, News, SaaS $ 1.350.000,00 01.01.2012
Cloud Computing, Cloud Security,
Computer Vision, Content
Delivery Network, Network
Azion Palo Alto, California Security 01.01.2011
Cloud Computing, Internet of
Things, IT Infrastructure,
Network Hardware, Software,
Rigado Portland, Oregon Wearables, Wireless $ 20.224.999,00 18.06.2010

Analytics, Big Data, Developer


Platform, Internet of Things,
Edge Intelligence Software Boston, Massachusetts Machine Learning $ 6.650.000,00 28.04.2010

Appendix 3: Automation Pyramid

The automation pyramid shows how


manufacturing firms traditionally connect
different levels internally. The pyramid,
however, excludes relevant actors outside the
firm, like suppliers or end-customers. The
problem within IIoT applications are, for
example, different computers with different
cloud connections that offer information and
data in different ways. In other words, the
Source: Own Creation
machines speak a different language.

XIII
Appendix 4: Elements of an autonomous driving system

This overview lists essential elements


within an autonomous car that were
investigated in a study by
McKinsey&Company in 2017. Edge
computing refers to all real-time data
processing that takes place inside the car.
One can see that the cloud plays an
important role in providing traffic data or
road maps.

Source: Heineke et al. (2017, p. 7)

XIV
Declaration of Authorship

I swear that I have written this thesis on my own and with no other help than the literature and other

supportive material listed in the appendix. Citations of sentences and parts of sentences are declared as such,

while other imitations are clearly marked and linked to original sources with regard to extent and intention

of the statements made. This thesis has never been handed in to any examination authority before and it is

also not yet published.

XV

You might also like