Michel Rauchs, Apolline Blandin, Kristina Klein,
Gina Pieters, Martino Recanatini, Bryan Zhang
December 2018
2ND GLOBAL CRYPTOASSET
BENCHMARKING STUDY
The Cambridge Centre for Alternative Finance (CCAF) is an international and interdisciplinary research centre based at the
University of Cambridge Judge Business School. It is dedicated to the study of innovative instruments, channels, and systems
emerging outside of traditional finance. This includes, among others, crowdfunding, marketplace lending, alternative credit
and investment analytics, alternative payment systems, cryptoassets, distributed ledger technology (e.g. blockchain) as well as
related regulations and regulatory innovations (e.g. sandboxes and RegTech).
TABLE OF CONTENTS
FOREWORD ......................................................................................................................................5
RESEARCH TEAM ............................................................................................................................6
ACKNOWLEDGMENTS ................................................................................................................7
EXECUTIVE SUMMARY ............................................................................................................ 10
METHODOLOGY .......................................................................................................................... 14
SETTING THE SCENE .................................................................................................................. 17
The Year in Review ..........................................................................................................................................................................17
SECTION 1:
THE CRYPTOASSET INDUSTRY ............................................................................................. 19
1.1 Segments .....................................................................................................................................................................................19
Industry Structure .............................................................................................................................................................................. 19
Mining Segment................................................................................................................................................................................... 20
Storage Segment ................................................................................................................................................................................. 21
Payments Segment ............................................................................................................................................................................ 24
Horizontal Expansion: The Growth of Multi-Segment Firms ...................................................................................... 25
1.2 Industry Growth .......................................................................................................................................................................26
1.3 Geography ..................................................................................................................................................................................28
A Global Industry ................................................................................................................................................................................ 28
Legal Headquarters and Operations ......................................................................................................................................... 28
SECTION 2: GLOBAL USAGE ................................................................................................... 30
2.2 Who Is Using Cryptoassets? ................................................................................................................................................31
Total Users ............................................................................................................................................................................................. 32
User Types .............................................................................................................................................................................................. 34
User Activity ......................................................................................................................................................................................... 35
User Location........................................................................................................................................................................................ 35
2.3 Cryptoasset Usage Characteristics ..................................................................................................................................37
On-chain Payments............................................................................................................................................................................ 37
Off-chain Payments ........................................................................................................................................................................... 39
Decentralised Applications and Timestamping ................................................................................................................... 40
Speculation and Investment .......................................................................................................................................................... 40
SECTION 3: GATEWAYS AND ECONOMIC CONNECTIONS .................................... 42
On-Ramps and Off-Ramps ............................................................................................................................................................. 43
Internal Cryptoasset Ecosystem Flows ................................................................................................................................... 45
Managing Volatility ........................................................................................................................................................................... 47
SECTION 4: STORAGE AND CUSTODY SEGMENT ...................................................... 49
Source Code .......................................................................................................................................................................................... 51
Key Storage Can Take Different Forms ................................................................................................................................... 52
Multi-Signature.................................................................................................................................................................................... 53
4
SECTION 5: REGULATIONS AND COMPLIANCE ........................................................... 54
5.1 The Impact of Regulations ....................................................................................................................................................54
User Impact of Regulations ........................................................................................................................................................... 54
Cryptoasset Firms Collaborate Directly with Regulators .............................................................................................. 56
5.2 KYC/AML Policies ....................................................................................................................................................................57
Implementation ................................................................................................................................................................................... 57
Criteria ..................................................................................................................................................................................................... 58
Account Suspensions and Closures ........................................................................................................................................... 59
5.3 Compliance Team .....................................................................................................................................................................60
5.4 Licensing ......................................................................................................................................................................................61
SECTION 6: IT SECURITY .......................................................................................................... 63
IT Security Team .................................................................................................................................................................................. 64
Security Audits .................................................................................................................................................................................65
Internal Policies ................................................................................................................................................................................... 67
SECTION 7: MINING SEGMENT ............................................................................................. 68
Cryptoasset Selection ...................................................................................................................................................................... 68
Influence on Decision-Making Process ................................................................................................................................... 70
Concentration Concerns ................................................................................................................................................................ 71
7.2 Hardware Manufacturing .....................................................................................................................................................72
Mining Equipment and Algorithms ............................................................................................................................................ 73
Distribution Channels ...................................................................................................................................................................... 75
How Concentrated Is Manufacturing? ..................................................................................................................................... 75
7.3 Mining Facilities .......................................................................................................................................................................77
Meet the Hashers ............................................................................................................................................................................... 77
Facility Set-up Decision Factors .................................................................................................................................................. 77
Distribution of Mining Facilities .................................................................................................................................................. 78
How Much Energy Does Cryptoasset Mining Consume? .............................................................................................. 81
How Wasteful Is Cryptoasset Mining? .................................................................................................................................... 83
What Do Miners Think? ................................................................................................................................................................... 85
7.4 Pool Operators ..........................................................................................................................................................................86
Pool Operations ................................................................................................................................................................................... 86
Pool Concentration ............................................................................................................................................................................ 87
FUTURE OUTLOOK ..................................................................................................................... 90
APPENDIX: SENTIMENT QUESTIONS ................................................................................ 94
5
FOREWORD
It is my great pleasure to announce the release of the second Global Cryptoasset
Benchmarking Study produced by the Cambridge Centre of Alternative Finance
based at the University of Cambridge Judge Business School. It examines
significant developments in the global cryptoasset ecosystem that have occurred
since the publication of our initial benchmarking study of cryptocurrencies in April
2017. The emphasis on 'global' in the title of this study is critically important given
the increasingly fluid, borderless nature of the cryptoasset industry. It also reflects
a core competence of our research centre, which is engaging in empirical research
investigating global technology-enabled financial innovation emerging outside of
the incumbent financial system. For our 2nd cryptoasset report the research team
spent several months collecting data from more than 180 entities in 47 different
countries, which represents a 25% increase in both the number of participants
and countries represented in comparison to our 2017 benchmarking report.
Our series of benchmarking studies analysing emerging forms of alternative
finance provides a comparative global snapshot of rapidly developing ecosystems
impacting the incumbent financial system. Our goal from the outset was that
these periodic reports would become a valuable reference for a wide audience
of actors in the financial system, including disruptive product and service
innovators, incumbent financial services firms, investors, academics, regulators
and policymakers, and the general public. Each of these constituents deserves
to be heard in debates about financial innovation, and few finance innovations
have been as controversial and attracted as much misinformed opinion as the
developments associated with cryptoassets. Our aim is to inform these voices
by providing empirically-based evidence of developments to provide common
points of reference to build upon. Sometimes this challenges prevailing wisdom.
For example, the analysis of excess renewable energy used by a share of mining
facilities suggests that the negative environmental externalities and associated
costs of the energy consumed by proof-of-work consensus systems could be
lower than previous estimates. We continue to believe that good research should
generate at least as many new questions as it answers, and we hope this report
passes that test.
Dr. Robert Wardrop
Director
Cambridge Centre for Alternative Finance
6
RESEARCH TEAM
Michel Rauchs: Michel is the Lead in Cryptocurrency and Blockchain at the Cambridge Centre for
Alternative Finance. He co-authored the inaugural benchmarking studies on the cryptoasset and
enterprise blockchain industries, and was the Project Lead of the Distributed Ledger Technology Systems:
A Conceptual Framework report.
m.rauchs@jbs.cam.ac.uk

@mrauchs
Apolline Blandin: Apolline is a Research Manager in Cryptocurrency and Blockchain at the Cambridge
Centre for Alternative Finance. Prior to joining CCAF, she graduated from Peking University and the
London School of Economics with a dual Master's degree in International Affairs. Her research has
mainly focused on mobile finance and financial inclusion in China.
a.blandin@jbs.cam.ac.uk

@ApollineBlandin
Kristina Klein: Kristina is a Visiting Student at the Cambridge Centre for Alternative Finance. She is
pursuing a Master's degree in Management and Technology at the Technical University of Munich (TUM)
and focuses on entrepreneurship and computer science.
k.klein@jbs.cam.ac.uk

@kklein93
Dr. Gina Pieters: Gina is a Lecturer at the Department of Economics at the University of Chicago and a
Research Fellow at the Cambridge Centre for Alternative Finance. Her research examines the economic
implications and behaviour of cryptocurrencies across different currencies and monetary systems.
gcpieters@uchicago.edu

@ProfPieters

Martino Recanatini: Martino is a Visiting Student at the Cambridge Centre for Alternative Finance. He
is pursuing a Master's degree in Finance and Banking at the Politecnica delle Marche University in Italy.
His Master's thesis assesses the potential impact of DLT systems on securities post-trading services.
m.recanatini@jbs.cam.ac.uk
@marecanatini
Bryan Zhang: Bryan is the Executive Director and a Co-Founder of the Cambridge Centre for
Alternative Finance. He has co-authored more than 20 reports on financial innovation and regulatory
innovation.
b.zhang@jbs.cam.ac.uk

@BryanZhangZ
7
ACKNOWLEDGMENTS
We would like to thank Liu Feng and the ChainNews team for providing a Chinese version of the surveys,
Miguel Klaggues from the Asociacin Bitcoin Chile for translating the surveys into Spanish, as well as
Kim Cheol Hwan and Seowon Park from the Korean Blockchain Industry Promotion Association (KBIPA)
for the Korean survey version. Nick Chong (Quoine) and Fiorella Velazquez (BitInka) provided helpful
comments and undertook significant efforts in helping distribute the surveys.
Special thanks go to Keith Bear (CCAF) and Kathryn Vagneur (CCAF) for providing invaluable feedback
in terms of survey design, data analysis, and report structure, Louise Smith for the beautiful design of the
report, as well as Derek Snow for his thorough review of an early draft. Our interns Hatim Hussain, Jaya
Lalwani, Jinjun Liu, Thomas Eisermann, Ouafaa Hmaddi, and Sabine Damborska deserve special thanks
for their tireless work and efforts in helping make this report happen. Finally, we would like to thank the
entire CCAF team and Kate Belger in particular for their continuous support and assistance.
We would also like to thank the following organisations for helping distribute the surveys to potential
respondents in their respective countries and regions:
Asociacin Bitcoin Chile, Associao Brasileira de Criptoeconomia (ABCripto), Association of Cryptocurrency
Enterprises and Startups Singapore (ACCESS Singapore), Bitcoin Argentina, Chaintech, Colombia Fintech,
Estonian Cryptocurrency Association, Ghana Blockchain Society, Israeli Blockchain Association, the Korean
Blockchain Industry Promotion Association (KBIPA), the Bitcoin Foundation, and the Nordic Blockchain
Association.
8
This research study would not have been possible without the generous support and participation
from industry actors: we would like to express our gratitude to the following cryptoasset entities for
contributing to this research study by completing our surveys. Some survey respondents prefer not to
publicly disclose their participation.
B E L E M
9
10
EXECUTIVE SUMMARY
Since the publication of the first Global Cryptocurrency Benchmarking Study in April 2017, the
cryptoasset ecosystem has undergone significant changes: the aggregate market capitalisation of
cryptoassets skyrocketed from $30 billion to more than $800 billion at its peak in early January 2018,
until coming down again to hover at around $200 billion.
The surge in prices and subsequent fluctuations was accompanied by growing interest and attention
from the general public and media, driving in new retail investors, speculators, and institutional investors.
The industry was confronted with massive inflows of new users and funds, a situation not all actors had
adequately prepared for. Growing interest from the institutional side contributed to the emergence of
custom services tailored to meet the needs and requirements of this new type of demand, leading to a
deeper interweaving of the industry and the incumbent financial system.
Between May and July 2018, the research team collected survey data from over 180 start-ups,
established companies, and individuals from 47 different countries across all major regions. The objective
of the study is to provide new insights into the current state of the ecosystem and, in combination with
publicly available data sources, capture major trends of the rapid market development. The analysis
focuses in particular on the following four key industry segments: mining, exchange, storage, and
payments.
The analysis reveals six main findings:
Millions of new users have entered the ecosystem, but most remain passive
Total user accounts at service providers now exceed 139 million with at least 35 million
identity-verified users, the latter growing nearly 4X in 2017 and doubling again in the first
three quarters of 2018. Only 38% of all users can be considered active, although definitions
and criteria of activity levels vary significantly across service providers.
Firms are increasingly operating across segments
The cross-segment expansion observed in 2017 has continued: 57% of cryptoasset service
providers are now operating across at least two market segments to provide integrated
services for their customers, compared to 31% in early 2017.
Multi-coin support is rapidly expanding
Multi-coin support has nearly doubled from 47% of all service providers in 2017 to 84% in
2018; a trend primarily driven by the emergence of common standards on some cryptoasset
platforms (e.g. ERC-20 on Ethereum) that has resulted in a rapid increase in the supply of
tokens.
The majority of identified mining facilities use some share of renewable energy sources as
part of their energy mix
The study estimates that as of mid-November 2018, the top-6 proof-of-work cryptoassets
collectively consume between 52 and 111 TWh of electricity per year. The mid-point of the
estimate (82 TWh) is the equivalent of the total energy consumed by the entire country of
Belgium but also constitutes less than 0.01% of the world's global energy production per
year. A notable share of the energy consumed by these facilities is supplied by renewable
energy sources in regions with excess capacity.
Mining is less concentrated than commonly perceived
Cryptoasset mining appears to be less concentrated geographically, in hashpower ownership,
and in manufacturer options than commonly depicted: the mining map exhibits that hashing
2nd Global Cryptoasset Benchmarking Study
11
facilities and pool operators are distributed globally, with growing operations in the USA and
Canada.
Self-regulatory efforts reflect growing industry maturity
Industry actors are pro-actively adopting measures that appear to comply with existing
regulation despite not necessarily being explicitly subject to regulations. The increasing number
of self-regulatory initiatives, combined with the emergence of sophisticated and professional
services, reflect the growing maturity of the industry.
Other notable findings include the following (ordered by section):
The Cryptoasset Industry
The industry has experienced substantial growth in terms of full-time equivalent (FTE)
employees: 2017 year-on-year growth rates reached 164%, driven primarily by the
exchange and storage segments.
Firm size has also increased significantly: the average firm now employs a median number of
20 staff, up from five employees in 2016.
While 21% of surveyed firms have their legal HQ in a different country than their
operational HQ, only 7% have their legal HQ in a different geographic region, suggesting
that while organisations may be willing to locate to nearby countries to exploit regulatory
arbitrage, many are not willing to move too far afield.
Global Usage

Individuals constitute the largest share of the user base (primarily served by exchanges and
multi-segment firms); payment service providers and storage providers have the highest
share of business users among service providers (26% and 32%, respectively).
Firms predominantly serve customers based in the region where they have their operational
HQ.
Both on-chain and off-chain transaction volumes have significantly increased in 2017;
behaviours consistent with speculation and long-term investment still account for the vast
majority of cryptoasset usage.
The share of high-value transactions (i.e. above $1,000) for cross-border payments
processed off-chain rose from 34% in 2016 to 46% in 2017, a trend that is mirrored by on-
chain transactions as well.
While Bitcoin's median on-chain transaction size has consistently grown since 2016, other
cryptoasset systems have declining median amounts per transaction.
Gateways and Economic Connections
The cryptoasset ecosystem is becoming more connected to traditional finance due to the
emergence and growth of gateways bridging both systems, as well as growing regulatory
clarity. The relatively small size of the industry in the global financial market poses no
systemic risk at this time.
Fiat-to-cryptoasset (and vice-versa) trades are allowed on some exchanges and payment
platforms, but not allowed on others. For fiat-supporting exchanges, these fiat-to-
cryptoasset trades make up the majority of trading volumes, demonstrating continuous
in- and outflows from the cryptoasset ecosystem to the incumbent financial system and the
real economy.
Bank wires dominate supported methods for both deposits and withdrawals; the use of
physical cash is more popular in Asia-Pacific than in other world regions.
EXECUTIVE SUMMARY
12
Service providers support a greater number of deposit options than withdrawal options,
suggesting that entering the ecosystem is generally easier than exiting.
69% of surveyed payment service providers have existing relationships with established
traditional payment networks, but difficulties of entering and maintaining good banking
relationships remain a primary concern, particularly for exchanges.
Storage and Custody Segment
Custody of cryptoassets is diverse: 62% of large entities retain control over customer funds
compared to only 30% of small firms. Similarly, firms operating across multiple segments
tend to take user funds into custody more often than companies specialised in one segment.
Two-thirds of specialised custodial exchanges do not have a refund procedure in the case of
customer funds getting lost or stolen.
The share of funds held in cold storage has slightly decreased over 2017 to enable quick on-
demand access, but is still above 80% of all funds.
Regulations and Compliance
Cryptoasset service providers are fostering their compliance efforts, even when not
explicitly subject to regulatory oversight: 37% of cryptoasset-only service providers have an
in-house compliance team and more than half perform KYC/AML checks.
While an average of 14% of KYC/AML checks result in service providers not opening new
accounts or closing existing accounts, some firms claim figures between 50% and 80% - well
above comparable traditional finance benchmarks.
The majority of surveyed companies rely on traditional services for third-party support in
conducting KYC/AML checks rather than specialised blockchain analytics providers.
Only 5% of surveyed cryptoasset-only service providers hold an operating license for their
jurisdiction, as opposed to 39% of fiat-supporting entities. However, 30% of cryptoasset-
only service providers are planning to apply for a license or register with local authorities,
which reflects the industry's willingness to proactively engage with compliance.
There are active efforts at industry self-regulation, with most of entities collaborating with
regulators and policymakers to address regulatory issues.
Changes in the regulatory environment have a measurable impact on operations: 38% of
fiat-supporting service providers have closed a location as a result of regulatory actions.
However, overall changes in the regulatory environment appear to have a greater impact on
encouraging location openings rather than causing closures.
IT Security
With more than $1.5 billion stolen from cryptoasset exchanges and storage providers alone
to date, IT security has become a crucial operational aspect: specialised storage providers
take the highest security precautions of all surveyed firms and dedicate the largest
headcount and budget share to IT security of all firms.
Providing regular training programmes for staff has become a common industry standard
as a substantial number of breaches have been caused by employee wrongdoing and/or
negligence.
We observe a lack of transparency on both external and internal security audits: more than
80% of firms do not publicly share information about security audits, indicating a general
unwillingness to divulge security-critical information.
2nd Global Cryptoasset Benchmarking Study
13
Mining Segment
Miners' concerns about the three main types of mining concentration (control over
hashpower, geographic distribution of hashpower, and the geographic distribution of
hardware manufacturing) have grown in 2017.
China remains in the top-3 countries to host mining farms; but the USA and Canada have
witnessed a rapid growth of mining farm openings over the past year, often associated with
the availability of cheap hydroelectric power.
Access to high-volume and low-cost electricity as well as stable political and friendly
regulatory environments are the major determining factors for hashers to choose an
operational location.
Over half of identified mining facilities, weighted by megawatts of electricity consumed,
have some share of renewable energy as part of their total energy mix. An increasing
number of hashing facilities are moving to regions with abundant low-cost electricity
generated by hydroelectric power.

Whilst many miners acknowledge the issue of environmental impact of PoW, most would
not advocate for switching to a new, less resource-intensive consensus algorithm.
The total number and geographic distribution of mining pools greatly varies from one
cryptoasset to another. While a third of surveyed pools are fully controlled by a single
person, past events show that low switching costs keep a check on operator behaviour.

A small share of pool members provides the majority of total pool hashpower: on average,
the top-10% of users contribute 68% of the pool's hashrate (top-1% contributes one third of
pool hashpower).
ASIC mining hardware manufacturing is dominated by a few producers; Ethash, SHA-256
and Equihash are the most supported mining algorithms.
Future Outlook
The trend towards increased multi-coin support is likely to continue: all single-coin storage
providers plan to support more cryptoassets in the near future.

Innovations in trust-minimised off-chain payment networks ("layer-2 solutions" such as
Bitcoin's Lightning Network) are thought to have the largest impact on service providers'
business models and operations.
Storage providers and multi-segment firms see stablecoins as a business-enhancing
opportunity, whereas non-fungible tokens (e.g. digital collectibles such as game items) are
generally thought to have a limited impact in the coming 12 months.
METHODOLOGY
14
METHODOLOGY
The Cambridge Centre for Alternative Finance carried out two online surveys between May and July
2018 via secure web-based questionnaires. The Cryptoasset Service Providers Survey was directed at
organisations operating in one or more segments of the cryptoasset industry as defined by our taxonomy
(specifically exchange, storage and payments), whereas the Cryptoasset Mining Survey targeted both
organisations and individuals involved in mining activities.1
The research team used various channels to disseminate the surveys globally in order to gather a
representative sample of the industry geographic dispersion. Both surveys were available in English,
Chinese, Spanish, and Korean. Surveys were distributed directly via email invitations to industry
contacts, as well as indirectly by sharing public links on social networks (e.g. Twitter, LinkedIn) and
Internet forums (e.g. Reddit, Bitcointalk). The research team also partnered with several national
cryptoasset associations to advance survey dissemination locally. The collected data was encrypted
and safely stored, accessible only to the authors of this study. All individual company-specific data was
anonymised and analysed in aggregate by industry segment, organisation size, supported assets, custody
types, and region.
Data was collected from more than 180 entities globally across 47 countries
For cases in which currently active major companies did not contribute to our study, the dataset
was supplemented with additional desktop research and web scraping using commonly applied
methodologies. Furthermore, publicly available data from a variety of sources was used to complement
survey data.
In total, survey data was collected from more than 180 entities across 47 countries and five world
regions: 127 firms participated in the Cryptoasset Service Providers Survey and 57 entities (22
organisations and 35 individuals) completed the Cryptoasset Mining Survey.2 Follow-up phone calls or
emails were used to clarify survey responses if needed, with further quality assurance provided by
comparing results to available public data if feasible.
We estimate that our benchmarking study captures more than 75% of the global
economic activity in the four cryptoasset industry segments covered in this report
Figure 1 provides a breakdown of survey participants by geographic region and firm age for each
dataset. Both samples reflect the global nature of the industry and incorporate a mix of new entrants
and incumbent firms. It is worth noting that the Bitcoin white paper was posted online just 10 years ago,
with the first large Bitcoin businesses established only 8 years ago: any firm 5 years or older has been in
existence for nearly the entirety of the life of the cryptoasset industry.
1
Other segments of the industry (e.g. Initial Coin Offerings/ICOs) are not covered by this analysis. All data points presented in the
following pages will be based on survey data, unless explicitly stated otherwise.
2
This represents an increase of 36 firms relative to the 2017 benchmarking survey. While the survey sample represents less than
25% of all identified entities in the four segments, we estimate that the study captures 75% of economic activity in the industry.
2nd Global Cryptoasset Benchmarking Study
15
Figure 1: The organisations in this study sample represent both industry veterans and new entrants
from around the world
Note: individual miners have been removed from the firm age calculation in the mining sample.
European companies dominate the service provider study sample, while individuals and mining
organisations based in Asia-Pacific take 40% of the mining study sample. Relative to last year, the survey
received more responses from each region, with firms from South America as well as the Middle East
and Africa (MEA) representing a small but growing share of respondents in both mining and service
providers.3 Some respondents expressed their reluctance to participate in the study because of changes
in their immediate regulatory environment.
3
We only provide geographic breakdown of results if it does not compromise survey anonymity.
Geographic
Distribution
Firm Age
Service Provider Study Sample
Asia Pacific
Europe
North America
Middle East and Africa
South America
New entrants
1 year
3 years
2 years
4 years
5+ years
9%
17%
6%
43%
24%
Mining Study Sample
12%
21%
2%
25%
40%
13%
14%
21%
22%
24%
7%
7%
27%
7%
7%
13%
40%
METHODOLOGY
16
Classification Terminology:
Large firms: entities with 40 or more full-time equivalent employees.4
Cryptoasset-only firms: entities that exclusively handle cryptoassets.
Fiat-supporting firms: entities that handle both cryptoassets and fiat currencies.
Multi-segment firms: entities that operate in multiple industry segments.
Custodians: entities that keep customer funds in custody.
Non-custodians: entities that do not keep control of customer funds.
4
An exception has been made for mining organisations, where FTE size is a less important indicator of relative economic influence
and position within the segment. Large and small miners are differentiated based on considerations of their hashpower,
importance to the ecosystem, or reputational prominence.
2nd Global Cryptoasset Benchmarking Study
17
SETTING THE SCENE
Focus on the Data Layer
The recently published study Distributed Ledger Technology Systems: A Conceptual Framework laid the
foundation for a comprehensive framework and terminology for the cryptoasset, blockchain and
distributed ledger technology (DLT) fields.5 It introduces a conceptual framework that divides a given
DLT system into a set of three layers: protocol, network, and data.
The primary focus of this study will be on the intermediaries (service providers and miners) that interact
with the data layer.6 This includes cryptocurrencies that play an essential role in the incentive design of
their respective DLT system as well as tokens that grant their holder with the right to access specific
functionalities of applications runnings on an existing DLT system.

From Cryptocurrencies to Cryptoassets
The astute reader may have noticed that this year's edition of the study uses the term
cryptoasset rather than cryptocurrency in the title. 2017 has seen a tremendous explosion
in the number of tokens that have been issued on top of existing platforms rather
than coming with their own, custom distributed ledger. This trend requires expanding
the vocabulary to move the discussion from cryptocurrencies to the broader term of
cryptoassets, which appears to have become the commonly-accepted umbrella term when
referring to the ensemble of (public) blockchain-based tokens, including cryptocurrencies.
The study will provide an empirical analysis of the four key cryptoasset industry segments (mining,
exchange, storage, and payments). The report will only sporadically touch on the underlying protocol and
network layers.
The Year in Review
During 2017, the total cryptoasset market capitalisation climbed from $18 billion in January to a
staggering $600 billion in December, raising questions about the cryptoasset market as a whole being
a giant bubble.7 The desire to be an early investor in "the next Bitcoin" further fueled speculative
investment. However, prices across the entire cryptoasset ecosystem started to tumble in January 2018,
moving downwards uniformly across all cryptoassets. Despite a few rebounds in early 2018, the price
decline has continued throughout the year and resulted in the evaporation of more than $600 billion of
market capitalisation.8
5
The study, authored by an interdisciplinary team of academics and industry experts, was published in September 2018: Rauchs,
M., Glidden, A., Gordon, B., Pieters, G., Recanatini, M., Rostand, F., Vagneur, K., and Zhang, B. (2018) Distributed Ledger Technology
Systems: A Conceptual Framework. Available at: https://www.jbs.cam.ac.uk/faculty-research/centres/alternative-finance/
publications/distributed-ledger-technology-systems/ [Accessed: 02 December 2018].
6
The data layer covers the nature and meaning of the final records produced by the DLT system. In the case of open, public, and
permissionless systems, these records primarily refer to the creation, transfer, and "destruction" of native cryptoassets.
7
The term "market bubble" generally refers to a situation where assets are traded at prices that substantially exceed their
fundamental value. In the case of cryptoassets, the definition of a fundamental value is both difficult and controversial to define
and determine.
8
Market capitalisation as a measure of network value is incomplete and relatively easy to manipulate. It thus bears the question
how much of the $600 billion in lost market capitalisation had been "real" gains in the first place.
SETTING THE SCENE
18
The rapid increase in Bitcoin prices spilled over to other cryptoassets and brought both sustained
media attention and new speculative investors (retail and institutional). The entry of traditional financial
services firms into the cryptoasset market and new offerings such as Bitcoin futures, specialised custody
solutions, and dedicated cryptoasset hedge funds further fuelled the expansion of the industry. However,
this also brought with it increased regulatory attention.
Tokens became more popular in the ecosystem as well, primarily driven by the wide adoption of the
ERC-20 standard on the Ethereum network and the resulting proliferation of Initial Coin Offerings
(ICOs). This led to a boom in token-based fundraising and a flurry of ICO activities globally. Blockchain
forks9 also became more common in 2017, further increasing the number of offerings in the cryptoasset
ecosystem by splitting existing coins into separate cryptoassets.10 The increase in interest and
subsequent usage brought into the foreground limitations of base layer scaling and led to the launch of
so-called "layer-2 solutions", such as the eagerly-awaited Lightning Network on Bitcoin.11
These developments have left a mark on industry actors: according to data collected from survey
participants in both 2017 and 2018, their views on various topics have changed considerably.
Particularly operational risks are perceived to have significantly increased: exchange operators consider
all listed risk factors more urgent in 2018 than the year before, whereas miners also tend to be faced
with increasing challenges (Tables 6 and 9 in Appendix).
9
A hard fork constitutes a controversial change to the protocol rules of a DLT system that causes the network to split into two
separate systems, each having their own cryptoasset.
10
Bitcoin alone had at least eleven known hard forks in 2017: Bitcoin Cash, Bitcoin Gold, Bitcoin Diamond, Super Bitcoin, Bitcoin
Platinum, Lightning Bitcoin, Bitcoin God, Bitcoin Uranium, Bitcoin Cash Plus, Bitcoin Silver, and Bitcoin Atom. Most of these forks
have seen negligible activity and adoption.
11
During the height of the boom, the Bitcoin blockchain experienced significant delays in processing transactions, with average fees
rising to levels above $50. Similarly, the Ethereum blockchain was clogged for a few days because of a single gaming application
that suddenly became popular (CryptoKitties). Layer-2 solutions refer to a variety of techniques that aim to materially increase
transaction speed and throughput as well as substantially decrease transaction costs by moving payments off-chain in a trust-
minimised manner.
2nd Global Cryptoasset Benchmarking Study
19
SECTION 1:
THE CRYPTOASSET INDUSTRY
1.1 Segments
Industry Structure
In the ten years since the publication of the Bitcoin whitepaper, an entire industry has evolved around
cryptoasset systems to build and maintain basic infrastructure as well as to facilitate the use of the
platforms and their assets. While there are several smaller segments comprising a great variety of
additional services such as blockchain analytics, data, and ICO services, this study will limit its focus to
four key industry segments: mining, storage, exchange, and payments (Figure 2).
Figure 2: The cryptoasset industry can be broken down into four key segments
Note: firms can operate in multiple segments.
Key Cryptoasset Industry Segments
Exchange
Storage
Mining
Providing a platform for the exchange
of one cryptoasset for another asset
Share of service providers
providing direct services
Enabling the secure management
of wallets storing cryptoassets
Receiving newly minted
cryptoasset units as a reward
for processing transactions
on the network
Facilitating the use of
cryptoassets for all types
of payments
61%
72%
Payments
49%
SECTION 1: THE CRYPTOASSET INDUSTRY
20
The study further distinguishes between direct and outsourced services. Most entities are providing
direct services in the exchange segment (72%), followed by storage (61%) and payments (49%) segment.
In some cases, service providers partner with a third party to outsource specific activities often those
that belong to a different segment. For instance, a storage provider may decide to partner with a third-
party exchange to offer in-wallet purchases and sales of cryptoassets. In such instance, the third party is
responsible for providing the exchange services and consequently for abiding by applicable regulations
as well. Among respondents, 12% of exchanges, 17% of storage providers, and 23% of payment service
providers are contracting out to an external party. The remainder of the report focuses on entities only
providing direct services.
Mining Segment
The mining segment comprises agents performing specific operations for the processing of public
blockchain transactions (Figure 3). During this process, new units of a specific cryptoasset can be
created.
Figure 3: Miners operate across a sophisticated value chain of distinct activities
The majority of mining organisations tend to specialise in a specific activity (46% of small miners and
56% of large miners). In contrast, a small number of large firms have pursued a continuous vertical
integration strategy that covers the entire value chain. Large firms tend to be older, whereas the majority
of individuals and half of small miners in the sample are new entrants.
Note: Section 7 covers the cryptoasset mining segment in greater detail.
The Mining Value Chain
Mining Hardware
Manufacturers
Entities designing and
building specialised
mining equipment
Proprietary Mining
Miners operating
mining equipment
on their own behalf
Hashing
Process of running mining
equipment to generate
hashes for finding a valid
proof-of-work
Cloud Mining
Services
Services renting out
hashpower generated
by own equipment to
customers
Pool Operators
Service combining
computational resources
from multiple hashers
and distributing
rewards
Remote Hosting
Services
Services hosting
and maintaining
customer-owned
mining equipment
2nd Global Cryptoasset Benchmarking Study
21
Storage Segment
Cryptoassets can be moved by signing transactions using private keys these keys are stored using
wallet softwares. Initially, wallets were simple software programs handling key management, but they
have evolved over time to support a variety of technical and commercial services. Many solutions provide
an easy-to-use interface for the end-user that abstracts away the complexity of key management.
Figure 4: Mobile wallets remain the most supported format; web wallet support has significantly
increased
Figure 4 shows the evolution of wallet options between 2017 and the second quarter of 2018. Mobile
and web wallets are the most widely offered storage formats, though cold-storage vault services have
gained in popularity in late 2017 with the influx of institutional investors, such as hedge funds. No
storage format has seen a decrease in support in 2018, indicating that the various forms of storage
options are not yet cannibalising each other.
Large storage providers support an average of three of the above types, compared to an average of two
types supported by small wallet providers. Storage-only service providers are more likely to specialise in
a particular activity, as opposed to multi-segment entities that often support multiple wallet formats.
Note: more detailed information on cryptoasset storage and custody is available in Section 4.
Mobile
Wallets
Desktop
Wallets
Web
Wallets
Tablet
Wallets
Vault
Services
Hardware
Wallets
Mobile wallets are smartphone applications that store
cryptoassets on mobile devices.
Web wallets are online applications that can be accessed
from any connected device via a browser.
Desktop wallets are key management software
programmes that run locally on a computer.
Tablet wallets are applications that enable users to store
cryptoassets on tablets.
Vault services provide sophisticated key management and
custody solutions combining multiple layers of security.
Hardware wallets are small devices that securely store
private keys without exposing them to connected machines.
Share of storage providers
supporting the listed formats
2018
2017
65%
38%
40%
24%
23%
62%
53%
42%
31%
31%
24%
SECTION 1: THE CRYPTOASSET INDUSTRY
22
Exchange Segment
Exchanges serve as on-off ramps for users to buy and sell cryptoassets - either in exchange for fiat
currency ("fiat-supporting"), another cryptoasset ("cryptoasset-only"), or other assets such as gold.
There are a variety of activities in the exchange segment that facilitate trade in different ways.
Figure 5: Brokerage services, order-book exchanges, and over-the-counter (OTC) trading desks are
the three major activities performed in the exchange segment.
Figure 5 shows that brokerage services, order-book exchanges, and OTC desks are offered by almost
half of all exchanges, with OTC desks greatly increasing in popularity since 2017. Exchanges exclusively
engages in a single activity are primarily providing order-book exchange services, brokerage services, or
OTC trading desks.
"Decentralised" exchanges continue to be a small, but growing, share of exchanges. While 7% of
respondents claimed to be decentralised - or P2P - exchanges, definitions of a decentralised exchange
vary widely: it is hence possible that the growth reflects a change in definition, rather than a true
increase in decentralised exchanges. Their rise in popularity may explain why surveyed P2P exchanges
are a lot more concerned about competition than centralised exchanges.
Brokerage
Services
OTC
Desks
Order-Book
Exchanges
Advanced
Trading
Services
"Decentralised"
Exchanges
ATMs
Brokerage services let users conveniently acquire
and/or sell cryptoassets at a given price.
Order-book exchanges are platforms that use a trading
engine to match buy and sell orders from users.
Over-the-counter (OTC) desks enable users to engage
in bilateral trades outside of formal trading venues in
order to avoid moving the market too much.
Advanced trading services allow users to buy portfolio
bundles and get access to more sophisticated tools such
as trading on margin.
"Decentralised" exchanges facilitate the trading of
cryptoassets without a central point of control. However, the
term is often loosely used to refer to different concepts.
Physical ATMs let users buy (and generally sell)
cryptoassets using physical cash.
Share of exchanges engaged
in the listed activities
No ATM service providers were surveyed
2018
2017
57%
57%
18%
41%
4%
49%
48%
45%
7%
31%
2nd Global Cryptoasset Benchmarking Study
23
Decentralised Exchanges (DEX) Rise in Popularity
Most exchanges are centralised businesses with dedicated operators, and therefore more
easily subject to regulations unlike direct peer-to-peer, blockchain-based trades. More
recently a number of "decentralised exchanges" (or DEXes) have emerged, promising
to provide a decentralised and trust-minimised alternative to traditional third-party
exchanges.
However, at present nearly all existing exchanges labelling themselves as DEX rely
on some degree of centralised control over certain exchange processes (e.g. order
matching) which raises questions whether they can claim to be "decentralised".12 Table 1
deconstructs exchange processes into three categories to classify the four exchange types.
Table 1: Decomposing exchange types - from custodial exchanges to DEXes
EXCHANGE
TYPE
1. CUSTODY
OF FUNDS
2. ORDER
MATCHING
3. CLEARING &
SETTLEMENT
a. Custodial
Exchange
Exchange
Exchange
b. Non-custodial Users
Exchange
Exchange
c. P2P
Users/
Exchange
Users/
Blockchain
Exchange
d. DEX
Users
Blockchain
Blockchain
Most self-labelled "DEXes" would be considered non-custodial exchanges (users retain full
control over their funds, but the exchange handles order matching and/or clearing and
settlement centrally) or P2P exchanges (primarily provide a flexible user matching platform
where users can decide whether to store funds at the exchange and perform the actual
trade outside of the platform).
In contrast, a true DEX uses a public blockchain for both order matching as well as clearing
and settlement while allowing users to maintain control of their funds for the entirety of
the trading process. While a DEX may pose a challenge for regulators, it is worth noting
that current implementations are cumbersome, expensive, and inefficient as they require
multiple on-chain transactions for a single exchange trade.
Of the surveyed exchanges, 41% are involved in only a single activity, while 59% are involved in two
activities or more. No respondent engages in more than four activities. Large exchanges tend to be
engaged in more activities than small exchanges (75% in two activities or more).
12
For a discussion on the multi-dimensional concept of "decentralisation", please see p. 44 of the Distributed Ledger Technology
Systems: A Conceptual Framework report. Available at: https://www.jbs.cam.ac.uk/faculty-research/centres/alternative-finance/
publications/distributed-ledger-technology-systems/ [Accessed: 02 December 2018].
SECTION 1: THE CRYPTOASSET INDUSTRY
24
Payments Segment
Payment service providers act as gateways to facilitate the use of cryptoassets for payments of all kinds
(Figure 6). To this end, the cryptoasset payments segment is composed of a variety of activities that
target different payment and user types.
Figure 6: Merchant services remain the most popular activity in the payments sector
Note: comparison with 2017 data is not possible given changes in activity classifications.
Large payment companies on average engage in three payment activities, while small providers engage
in two. Merchant services are the most popular payment activity, increasing slightly from 52% in 2017
to 57% in 2018. Cross-border flows (personal remittances and B2B transfers) are the next two most
popular services, with the number of payment providers supporting B2B payments more than doubling
from 19% in 2017 to 41% in 2018.
Merchant
Services
B2B Cross-
border
Personal
Remittances
Micropayments
General-
purpose
Platforms
Bill Payment
Services
Consumer
Payments
M2M
Payments
Merchant services process payments on behalf of cryptoasset-accepting
merchants.
Personal remittances generally refer to cross-border money transfers initiated
by workers abroad to their home country.
B2B cross-border payment platforms facilitate international transfers of funds
between businesses.
Micropayments are financial transactions that involve the transfer of very small
amounts of money.
General-purpose payment platforms provide an integrated platform offering a
variety of money transfer and payment services.
Bill payment services facilitate the use of cryptoassets for the payment of bills
(e.g. utilities, salaries, taxes).
Consumer payments services facilitate the use of cryptoassets as a payment
method for consumer purchases.
M2M payments constitute financial transactions occurring directly between two
machines.
Share of payment
service providers
engaged in the
listed activities
2018
57%
46%
41%
33%
30%
26%
28%
11%
2nd Global Cryptoasset Benchmarking Study
25
Horizontal Expansion: The Growth of Multi-Segment Firms
The lines between exchange, storage, and payments service segments have become increasingly
blurred: in 2017, 69% of surveyed service providers (excluding miners) were active in a single segment,
as opposed to 43% of service providers in 2018 (Figure 7). This change has come from firms expanding
their offerings across the different segments, and usually not from mergers between two horizontal
firms.
Figure 7: The shift towards multi-segment operations continues - more than half of service
providers are operating across two or more segments
Note: this chart does not include mining companies, as these - with a few exceptions - generally tend to focus exclusively on the
mining segment.
Large firms are more likely to operate across two or more segments, with 69% of large firms reporting
they do, compared with approximately half of small firms. The survey has insufficient data to determine
whether the firms are large because they are engaging with multiple segments, or whether reaching
a certain size grants the firm sufficient resources to operate across different segments and diversify
activities. Similarly, it is unclear whether firms are attempting to becoming a "one-stop shop" for
cryptoasset users, or whether the expansion is a move to provide more revenue streams to survive
frequent industry downturns.
Share of Service Providers Providing Direct Services
3+ segments
2 segments
1 segment
2017
12%
19%
69%
2018
28%
29%
43%
SECTION 1: THE CRYPTOASSET INDUSTRY
26
1.2 Industry Growth
The evolution of the total number of full-time equivalent (FTE) employees represents a good proxy for
assessing total industry growth. According to Figure 8, total industry headcount has grown by 164% in
2017, an expansion primarily driven by the exchange and storage segments.
Quarterly FTE staff growth in 2018 Q1 amounted to a staggering 78%: the accumulating backlog of
customer onboarding and support requests, the need for compliance staff to navigate the complex
landscape, engineers to make the platforms fit for drastic increase in traffic and usage, as well as
salespeople to onboard new clients are potential drivers of the segment job growth.
Figure 8: Industry growth in 2017 appears to be primarily driven by the exchange and storage
segments
Full-Time Equivalent (FTE) Staff
Total Growth Rate
Industry-wide
Segment Breakdown
Quarterly Growth
2018 Q1
YoY Growth
2017
78%
164%
Exchange-only
Exchange-only
Storage-only
Storage-only
Payment-only
Payment-only
Mining
Mining
Multi-segment
Multi-segment
82%
333%
79%
244%
72%
103%
67%
103%
83%
176%
2nd Global Cryptoasset Benchmarking Study
27
Organisations, including miners, more than doubled in size during the boom year of 2017 (Figure
9). Despite the global market crash at the beginning of the first quarter of 2018, firms in the sample
have continued to grow, with the median firm size increasing to 20 employees working exclusively on
cryptoassets, up from 13 at the end of 2017 (54% quarterly FTE growth). The job boom resulted in a
reported shortage of talent: the majority of surveyed exchanges reported concerns over good access
to talent. While the survey results were collected in May - five months after the price collapse - it is
possible that the sustained nature of the market crash has resulted in layoffs that are only beginning to
be realised now.
Figure 9: Organisations have also considerably grown in size - the typical miner has experienced
the highest growth in 2017
Increasing prices in the second half of 2017 meant that cryptoasset mining became more profitable for
a certain period before the difficulty adjustments kicked in: higher profitability resulted in mining firms
growing the most in size over 2017 (157% year-on-year growth), then slowing considerably in 2018 Q1.
Instead, firms specialising exclusively in the exchange segment took the lead in the first quarter of 2018,
with their median staff number growing by 82%.13 Service providers operating in multiple segments have
become the largest firms in terms of total employee headcount since 2017 and experienced the second-
highest FTE staff growth rates, with 65% year-on-year growth in 2017 and a quarterly growth of 150%
in 2018 Q1.
FTE figures widely vary between companies: the number of employees can range from one person
to several hundred, with one company employing more than 2,500 staff. No major differences can be
observed in terms of FTE growth rates between incumbent firms and newly-entered firms.
Overall, the market appears to be stable to price shocks; indicating that while cryptoassets may be
volatile, the economic employment generated by the cryptoasset industry is not. However, it may also be
too soon for the effect of the price decline on employment to be reflected in the data.
13
Figures are significantly lower for other segments, with 17% and 33% FTE staff growth for storage-only and payment-only
companies, respectively.
Full-Time Equivalent Staff
Firm Median
20
13
5
Exchange-only
20
11
5
Storage-only
14
12
7
Payment-only
8
6
3
Mining
17
18
7
Multi-segment
33
20
8
2016 YE
2017 YE
2018 Q1
2018 Q1
2017 YE
2016 YE
Quarterly
Growth
54%
YoY
Growth
160%
Industry-wide
Segment Breakdown
SECTION 1: THE CRYPTOASSET INDUSTRY
28
1.3 Geography
A Global Industry
The research team gathered supplementary data on headquarters of 561 companies active in at least
one of the four major cryptoasset industry segments. Figure 10 shows that the distribution of the global
cryptoasset industry is geographically uneven with high concentration of headquartered entities in
countries such as the USA, China, Japan, India, Canada, and the UK.
Figure 10: While global in nature, the cryptoasset industry is primarily driven by companies based
in North America, China, India, and Western Europe
Note: this map is based on operational HQ data from 561 companies active in at least one of the four major cryptoasset industry
segments. Colours represent the total number of identified entities from each country; stripes indicate when a country is also
covered by our study sample.
Legal Headquarters and Operations
Given the inherent cross-border nature of cryptoasset market activities, it is not uncommon for
cryptoasset entities to have operating activities in one country whilst being legally registered in another.
The average large entity maintains offices in three countries, compared to an average of two countries
for smaller entities. Multi-segment firms own and operate offices and facilities in more countries on
average than specialised companies.
The majority of surveyed companies have their operational headquarters and legal headquarters in the
same country: across the sample, slightly less than a quarter of firms (21%) have their legal HQ based in a
different country to their operating HQ. These results vary across industry segments: 22% of surveyed
service providers have their legal HQ in a different country, while only 14% of mining companies have a
legal HQ in a different country from their main operations.
Geographic Distribution of the Cryptoasset Industry
Number of identified entities
1-2
3-5
6-10
11-20
21-50
>50
Countries covered by survey sample
Canada
Brazil
India
China
Russia
Australia
United States
2nd Global Cryptoasset Benchmarking Study
29
Even if companies locate part of their activities in different countries, they may not necessarily move to
a new geographic region. Only 12% of service providers have their legal HQ in a different geographic
region; a situation that is even less common for surveyed mining companies (9%). Entities registered
in Europe appear to be 'regional', with facilities and offices being mainly based in Europe. Companies
registered in North America are 'cross-regional', with offices and facilities often located in multiple
regions.
The majority of industry actors have both operational and legal headquarters in the same country
This finding suggests that cryptoasset entities are more 'grounded' than previously thought: they are
more likely to operate in countries where they are legally registered - perhaps due to greater familiarity
with local regulatory structures - instead of looking for the most permissive legal environment. However,
focused regulatory attention increased during the latter half of 2017, therefore "regulatory arbitrage" of
legal and operational headquarters might be an emerging trend within the industry that would not have
been fully captured by the present study.
Note: more detailed information on regulation and compliance is available in Section 5.
SECTION 2: GLOBAL USAGE
30
SECTION 2: GLOBAL USAGE
2.1 Towards a Multi-Coin Universe
2017 saw explosive growth in the total number of cryptoassets. The selected major non-Bitcoin
cryptoassets all received increased support among surveyed service providers over 2018 (Figure 11).
The slight decline in bitcoin (BTC) support can be explained by the emergence of dedicated services
for specific cryptoassets. Privacy-focused coins such as zcash (ZEC) and monero (XMR) have grown in
popularity, yet still seem to occupy a niche.
Figure 11: Service provider support for the major cryptoassets has broadened
We observe differences between industry segments: for instance, bitcoin cash (BCH) is supported
by 77% of storage-only providers, but only 18% of payment-only companies and 22% of specialised
exchanges. This fact highlights that many cryptoassets have not yet been adopted into broader use (e.g.
consumer and cross-border payments).
Several interpretations co-exist to explain the increase in number of supported coins: the rise of ERC-
20 tokens, fuelled by ICOs and their subsequent listing on exchanges, as well as the growing number
of blockchain forks and airdrops are possible supply-side drivers. Meanwhile, tokens catering to niche
interests as well as speculation on price appreciation rather than native application usage increase the
demand for multiple coins.
Supported Cryptoassets
Share of Service Providers (Excl. Miners)
2018
2017
Litecoin | LTC
Ether | ETH
Bitcoin | BTC
92%
79%
61%
Bitcoin Cash | BCH
47%
Zcash | ZEC
22%
98%
33%
26%
Ripple | XRP
41%
13%
Monero | XMR
13%
8%
Stablecoins
11%
5%
Other cryptoassets
45%
24%
2nd Global Cryptoasset Benchmarking Study
31
Figure 12: Multi-coin support has nearly doubled from 47% in 2017 to 84% in 2018 across the
industry
In contrast to 2017, the vast majority of service providers are multi-coin-oriented today (Figure 12).
Surveyed storage providers have considerably expanded their cryptoasset coverage since 2017:
wallets with multi-coin support surged from 46% in 2017 to 90% in 2018, with 60% of wallets currently
supporting more than 3 cryptoassets as opposed to only 10% in 2017. Notably, all storage-only service
providers surveyed in 2018 exclusively support cryptoassets (i.e "cryptoasset-only").
An interesting observation is that companies specialised in payment services generally support fewer
cryptoassets than firms active in other segments: the majority of payment-only service providers merely
support one or two cryptoassets. In contrast, specialised exchanges support on average the broadest
number of cryptoassets.
2.2 Who Is Using Cryptoassets?
Information on the numbers and characteristics of cryptoasset users is sparse, therefore the results from
this survey represent a rare opportunity to gain information at the global level. A few central banks and
research institutes have issued surveys trying to determine the number and activity level of cryptoasset
users within their area of operation, but information is typically difficult to glean.
Cryptoasset users do not need to establish an account with a service provider in order to access and
use the underlying blockchain payment systems, and until recently most entities did not require users
to prove identities.14 Similarly, many users use a variety of different service providers and can thus have
multiple accounts. In some cases, users try to obfuscate their location using tools such as VPN servers
and the TOR network.
These factors make it difficult to provide precise estimates about the number, nature, origin, and activity
levels of cryptoasset users. The following results were constructed as conservative estimates of the
global cryptoasset user base, extrapolating from survey responses and publicly available data.
14
However, a growing number of service providers now require that users undergo Know Your Customer (KYC) and Anti-Money
Laundering (AML) checks (see Section 5 for more information), while certain jurisdictions such as South Korea and Japan require
the users to open an account at the bank the exchange has partnered with. We will refer to these users as "ID-verified users".
Multi-coin Support
2017
2018
Exchange
segment
Storage
segment
Payment
segment
89%
90%
77%
35%
61%
46%
SECTION 2: GLOBAL USAGE
32
Total Users
Most attempts to estimate the total number of users are piecemeal. A Bank of Canada survey in July
2018 estimated that the share of the Canadian population owning cryptoassets has nearly doubled from
2.9% in late 2016 to 5% in 2017 the equivalent of roughly 2 million users.15 Other studies estimate
that between 2% and 8% of US citizens have owned cryptoassets between the fall of 2017 and the first
quarter of 2018, which would amount to between 6 and 26 million users.16
Studies in other regions report similar findings: the Japanese Financial Services Agency (FSA) estimated
3.5 million domestic cryptoasset users (roughly 3% of the population) as of March 2018,17 whereas an
ING study published in June 2018 found that on average 9% of individuals in Europe own cryptoassets,
although significant differences between countries exist.18
Other studies set exclusively in Europe, Canada, Japan,
or the USA estimate 2-9% of the population hold cryptoassets
A naive extrapolation would then suggest that between 2% and 9% of the population of developed
countries would have owned cryptoassets as of mid-2018, a figure that would amount to between 28
and 126 million unique users.19 It should be noted that this extrapolation does not include cryptoasset
owners and users from developing countries, whose numbers are likely in the millions as well. However,
this does not necessarily imply that the actual number of cryptoasset owners worldwide substantially
exceeds the 100-million mark.20 Instead, these estimates should merely serve as an indication of the
potential number of users.
User Accounts Are Not (Necessarily) Equivalent to Users
The number of users and the number of user accounts are frequently conflated, even
though they represent very different notions. An individual user can hold multiple accounts
at any given service provider. Importantly, if a firm does not collect information on users,
it may not be able to link the multiple accounts to the user if a new identity, for example
a different email address, is associated with the account. Additionally, users can hold
accounts at more than one service provider. It follows that there are more user accounts
than users, more users than ID-verified users, and more ID-verified users than unique users
who engage in the cryptoasset ecosystem.
15
Bank of Canada (2017) Bitcoin Awareness and Usage in Canada. Available at: https://www.bankofcanada.ca/wp-content/
uploads/2017/12/swp2017-56.pdf [Accessed: 28 November 2018]; and Bank of Canada (2018) Bitcoin Awareness and Usage in
Canada: An Update. Available at: https://www.bankofcanada.ca/wp-content/uploads/2018/07/san2018-23.pdf [Accessed: 28
November 2018].
16
Blockchain Capital (2017) Bitcoin Survey Fall 2017. Available at: http://www.survey.blockchain.capital/ [Accessed: 28 November
2018]; and Finder.com (2018) Why haven't we all bought cryptocurrency yet? Available at: https://www.finder.com/why-people-
arent-buying-cryptocurrency [Accessed: 28 November 2018].
17
FSA (2018) . Available at: https://www.fsa.go.jp/news/30/singi/20180410-3.pdf [Accessed: 28
November 2018]
18
ING (2018) Cracking the Code on Cryptocurrency: Bitcoin buy-in across Europe, the USA and Australia. Available at: https://think.ing.
com/uploads/reports/ING_International_Survey_Mobile_Banking_2018.pdf [Accessed: 28 November 2018].
19
For the purpose of this analysis, we consider a country to be developed if it has a very high Human Development Index (HDI) score
(i.e. equal to or above 0.8). This currently applies to 59 countries worldwide with an aggregate population of around 1.4 billion
people.
20
Survey data can suffer from sample bias and be restricted to a certain geography whose context is fundamentally different than
another geography, among others. As a result, simple extrapolations should always be met with a healthy dose of scepticism.
2nd Global Cryptoasset Benchmarking Study
33
While it can be difficult to distinguish user accounts from individual users, survey data indicates that
there has been a substantial increase in the proportion of ID-verified users as a share of total user
accounts at service providers: while a mere 10% of total accounts could be attributed to ID-verified
users in 2016, one in every four accounts has been verified as of 2018 Q1.
There has been a substantial increase in the share of ID-verified
accounts: from 10% of user accounts in 2016 to 25% in 2018 Q1
Combining public data and survey findings, we estimate that the total number of user accounts at service
providers amounts to at least 139 million in late 2018 (Figure 13).21 Using a combination of verified user
data and the average share of ID-verified accounts described above, we also estimate there are currently
at least 35 million ID-verified users globally.22 While several limitations of the analysis above need to
be taken into account, we believe that the figures represent the lower-bound of the global cryptoasset
unique user base.23
Figure 13: Service providers currently serve at least 35 million ID-verified cryptoasset users, and
the industry is still undergoing rapid growth
The analysis reveals the continued growth in the number of user accounts at service providers
throughout 2017 and 2018. It also shows that KYC'ed user growth has dwarfed total user account
growth, which means that new users are more likely to get immediately verified. Growth rates were at
their highest in 2017, and the number of new user accounts as well as ID-verified users continued to
rapidly grow in 2018 as well.
21
The research team collected longitudinal account and user data of both small and larger service providers from publicly available
data sources such as press releases, news articles, company websites, social media, and public forums. This dataset was combined
with survey data from participating platforms and projects from 47 countries.
22
When available, absolute figures of ID-verified users supplied by survey participants were aggregated for each period. The ratio of
ID-verified users as a share of total accounts, calculated using survey data for each period, was then applied to the remaining total
account figures. The resulting figures were eventually added together to provide an estimate of the number of ID-verified users in
the ecosystem.
23
The analysis does not capture all accounts at service providers since no data was available for some major platforms (e.g. in
China) or individuals who do not use service providers. Together, these would contribute to an underestimation of total users.
On the other hand, there are no easy means to identify users with accounts at multiple service providers a practice that
would contribute to an overestimation. Overall, there are reasons to believe that the underestimation factors outweigh the
overestimation factors, which suggests that the current figure is a conservative lower-bound estimate.
Lower-bound Estimate of Total Cryptoasset Users
Total accounts
Total ID-verified users
20
0
40
60
80
100
140
120
160
Million2016 YE
45
5
2017 YE
85
18
2018 Q1-Q3
139
35
90%
63%
298%
94%
SECTION 2: GLOBAL USAGE
34
Table 2 removes the effect of the many new entrants by only analysing the account data of firms that
have been operating since 2016. All incumbent firms reported growth in the number of user accounts
though there is a consistently wide range of growth reported (from 24% to 2,900%). Given this reported
range, it is not surprising that the median growth is considerably lower than the average growth, but it is
still very robust.
At least 139 million user accounts have been created at cryptoasset
service providers, representing a minimum of 35 million ID-verified users
Consistent with the growth rates of the total accounts, firms saw higher growth rates in 2017 than 2018,
and higher growth rates of ID-verified users. This indicates that increased use of KYC and AML methods
in the industry also occur among established service providers who already have a substantial user base,
rather than just being the result of new entrants.
Table 2: Incumbent firms have experienced massive growth in user accounts and verified users in
2017
GROWTH RATES: AVERAGE (MEDIAN)
2017 YOY
2018 Q1-Q3
Average firm
Accounts
535% (233%)
161% (50%)
ID-verified users
977% (446%)
202% (79%)
Observed range
Accounts
24% - 2,900%
4% - 2,400%
ID-verified users
25% - 6,200%
19% -2,876%
Note: in order to avoid new entrant bias, this analysis only considers firms that have been active for the entire period of 2016-
2018. Outliers reporting over 3,000% growth in accounts or more than 7,000% for ID-verified users have been removed.
User Types
Conforming with popular narratives, survey data indicates that the majority of users both established
as well as new entrants are individuals and not business clients (Figure 14). Individuals can be
hobbyists, retail investors, consumers, or users seeking a better investment or payment alternative.
However, business clients comprise over one quarter of the users for payment-only and storage-only
services, suggesting that these two segments in particular represent opportunities to provide specialised
services specifically tailored for enterprises.
Figure 14: Individuals still constitute the majority of the user base of most service providers
Share of User Types
Exchange-only
Multi-segment
Storage-only
Payment-only
16%
22%
32%
46%
26%
58%
70%
78%
11%
6%
16%
19%
Individuals
Business clients
N/A
2nd Global Cryptoasset Benchmarking Study
35
The majority of business clients are comprised of cryptoasset hedge funds and online merchants
Multi-segment firms provide a fully-integrated suite of services that are especially popular with
individuals. Among business clients, 93% of surveyed entities primarily serve hedge funds (on average
60% crypto-focused hedge funds and 40% traditional hedge funds). The second most commonly served
category is merchants (86%), of which on average 63% are online merchants and 37% are brick and
mortar merchants. A further 35% and 30% of service providers report venture capital firms and other
institutional investors as business clients, respectively. Finally, 45% of service providers indicate having
miners as business clients, whereas 30% report serving other types of cryptoasset companies.
User Activity
While some use cases (e.g. long-term investment) do not require a user to actively use a service,
others such as short-term trading and frequent payments do. However, comparing active users across
platforms and services is difficult because the definitions of activity vary widelyeven within segments.
Weekly logins are the most popular criteria to determine user activity levels (supported by 28% of
service providers), followed by monthly measures related to activities that can involve simple logins,
deposits, trading, and withdrawals (26%).
The criteria for determining the level of user activity
vary significantly from one service provider to another
Using their own definitions, service providers report an increase in active users: an average of 35% of
users were considered active in 2016, 37% of users were active in 2017, and 38% of users in 2018 Q1.
However, the median service provider saw a decrease in user activity in 2017: 37% in 2016, falling to
35% in 2017, before rising to 40% in 2018 Q1. This difference reflects the range reported by firms: some
report as few as 5% active users, while the upper bound ranges from 85% (2016) to 91% (2017) and
80% (2018 Q1). The extent to which these ranges reflect different definitions instead of different levels
of user engagement is something this report cannot answer. Standardising definitions across activities
and segments would be a first step towards conducting the analysis described above.
User Location
Cryptoassets directly transacted via a public blockchain can easily move globally, but cryptoasset
transactions that move through service providers may be more region-locked. We find that entities
predominately serve customers based in the region of their operational headquarters: over half of
customers are based within the same region (Figure 15). North American and Asian-Pacific firms have
the largest share of non-domestic region customers, although on average a fifth of customers have
unknown locations. Middle Eastern and African firms have a quarter of their customer base in Europe.
European entities serve Asia-Pacific and North America customers in almost equal proportion (12% and
11%, respectively).
SECTION 2: GLOBAL USAGE
36
Figure 15: Firms predominantly serve regional customers based in the region where they have their
operational HQ
This analysis can be complemented with publicly available data sources. For instance, Coinmap shows
the geographic distribution of more than 13,000 businesses and merchants worldwide accepting
cryptoassets, which is dominated by North America, Europe, as well as some regions in South America,
South-East Asia, Southern Africa, and Oceania.24 The geographic distribution of Bitcoin ATMs reveals
a similar picture: according to Coin ATM Radar, more than 4,000 ATMs have been set up in over 70
countries, with the USA, Canada, Austria and the UK clearly dominating in absolute numbers.25
A recent analysis on transaction volumes at P2P exchange LocalBitcoins conducted by news and research
firm The Block revealed that when weighted by population size, usage was most popular in developing
countries that have lived through severe monetary turmoil (e.g. Russia, Venezuela).26 This may also
indicate a lack of reliable local exchange infrastructure in developing countries.
Overall, cryptoasset usage is a global phenomenon that involves users from all around the world. While
some regions (e.g. North America, Central and Eastern Europe, South-East Asia, and parts of South
America) and specific countries (e.g. USA, Canada, Japan, South Korea, China, UK, India, France) seem to
dominate in terms of active usage, other regions are catching up.
24
Data available at https://coinmap.org/#/world/13.58192090/13.35937500/2 [Accessed: 02 December 2018].
25
Data available at: https://coinatmradar.com/ [Accessed: 02 December 2018].
26
Cermak, L. (2018) Analysis: Russia and Venezuela dominating LocalBitcoins volumes, an intriguing proxy for bitcoin demand.
The Block. Available at: https://www.theblockcrypto.com/2018/10/26/analysis-russia-and-venezuela-dominating-localbitcoins-
volumes-an-intriguing-proxy-for-bitcoin-demand/ [Accessed: 02 December 2018].
Firm Operational HQCustomer locationAsia-
Pacific
Asia-
Pacific
South
America
78%
50%
50%
20%
13%
71%
28%
56%
20%
12%
South
America
Unknown
North
America
North
America
Middle
East and
Africa
Middle
East and
Africa
Europe
Europe
2nd Global Cryptoasset Benchmarking Study
37
2.3 Cryptoasset Usage Characteristics
Cryptoasset systems have integrated payment networks that enable the transfer of value between
users. The open and transparent nature of these systems allows the analysis and comparison of on-
chain transactions of the major cryptoassets, while our survey allows insights into off-chain payments
that are processed in internal database systems operated by service providers. However, it is difficult to
determine what a given cryptoasset transaction (or payment) is used for given that actual usage is often
entirely contextual and depends on a variety of factors (e.g. geographic location, access to alternatives,
etc.).
On-chain and off-chain transactions
On-chain: transactions that clear and settle directly on the respective blockchain.
Off-chain:


Trusted: transactions recorded by, and reliant upon, service providers for internal
clearing and settlement.


Trust-minimised: transactions based on payment channels using the blockchain
exclusively for settlement.


On-chain Payments
Figure 16: Bitcoin clearly dominates in terms of monthly on-chain transaction volumes, but loses
ground to Ethereum in the number of processed payments
On-chain Transaction Volumes and Number of Payments
$2
400,000
200,000
0
$4
300,000
100,000
$6
600,000
$8
700,000
$10
800,000
$0
Daily average on-chain transaction volume ($billion)Average number of daily payments2016 Q1
2016 Q2
2016 Q3
2016 Q4
2017 Q1
2017 Q2
2017 Q3
2017Q4
2018 Q1
2018 Q2
2018 Q3
2018 Q4
500,000
Bitcoin (BTC)
Litecoin (LTC)
Zcash (ZEC)
Ethereum (ETH)
Bitcoin Cash (BCH)
BTC - Volume (lhs)
ETH - Volume (lhs)
LTC - Volume (lhs)
BCH - Volume (lhs)
ZEC - Volume (lhs)
BTC - Payments (rhs)
ETH - Payments (rhs)
LTC - Payments (rhs)
BCH - Payments (rhs)
ZEC - Payments (rhs)
SECTION 2: GLOBAL USAGE
38
Figure 16 shows the on-chain transaction volume (measured in USD) and the average number of daily
payments of five public blockchain systems. Bitcoin (BTC) has the greater transaction volume, followed
by Ethereum (ETH). Bitcoin Cash (BCH), Litecoin (LTC) and Zcash (ZEC), which are usually small in
comparison. Transaction volume increased during the price boom, but subsequently decreased back to
the levels of early 2017. The number of payments for each coin is slightly different: specifically, there was
a brief period during which Ethereum surpassed Bitcoin in terms of processing the largest number of
payments. This implies that Ethereum payments tend to be of lower value than Bitcoin's.
The Difference Between Transactions and Payments
A general measure of blockchain throughput is the number of transactions that
cryptoasset systems can process over a predefined period. However, one transaction may
contain a bundle of several payments funds moving from different senders to different
recipients. This makes it possible for transactions and payments to move differently (e.g.
a decreasing number of transactions can support an increasing number of payments), and
therefore both must be analysed to obtain a full picture of a blockchain's use as a payment
system.
The total value of daily transactions processed by the top-5 cryptoassets has grown nine-fold during
2017 and are still showing positive growth of 44% in 2018 despite the cool-down. However, the median
transaction volume and transaction size has evolved differently across cryptoassets (Figure 17).
Figure 17: Bitcoin's on-chain transaction volume and transaction size have been continuously
growing since 2016, contrary to other cryptoassets
Comparing Transaction Volumes to Sizes
$0
Average daily transaction volume (on-chain)
Million
$500
$1,000
$1,500
$2,000
$2,500
$3,000
$3,500
$0
$300
$600
$100
$400
$700
$200
$500
$800
$900
Median transaction size (on-chain)2016
2016
2016
2017
2017
2017
2017
2018
2018
2018
2018
2018
Zcash (ZEC)
Bitcoin Cash (BCH)
Litecoin (LTC)
Bitcoin (BTC)
Ethereum (ETH)
2017
2nd Global Cryptoasset Benchmarking Study
39
Bitcoin has seen increasing transaction value and transaction size from 2016 to 2017, the only one of
the five represented cryptoassets to do so.27 Ethereum has seen growing median transaction volume
but constant low transaction size, indicating that it is used mostly for small-value transactions. This
is consistent with its status as a major cryptoasset platform for supporting and running applications
("dApps"). While most cryptoasset systems saw a decline in median transaction value during 2018,
Bitcoin Cash is unusual in the large decline in transaction value it experienced between 2017 and 2018.
Off-chain Payments
The speed and costs of many cryptoasset systems has resulted in a portion of payments being taken off-
chain, often facilitated by custodial service providers rather than trust-minimised second-layer solutions.
Payment service providers aggregate all transfers in their internal computer system and only use the
blockchain as a settlement layer for netting the outstanding transfers.28 This removes significant burden
from the underlying blockchain network by freeing up capacity, at the expense of putting users at the
mercy of service providers (e.g. decision to freeze accounts or block payments).
Figure 18: The share of high-value off-chain payments especially for cross-border transfers has
significantly increased
Survey data gives an indication of the nature of these payments (Figure 18). Transactions that are
domestic behave differently than transactions that are cross-border, implying that users employ
cryptoassets for different purposes in these two contexts. Domestic transactions are small, with
approximately half of all transactions under $100.
27
The transaction volume is the aggregate dollar-value amount processed by the system during a specific period, whereas
transaction size is the median transaction amount processed by the system during the same period.
28
Internal transfers do not actually move cryptoassets but consist in updating balances denominated in cryptoassets held in custody
by the service provider.
Transaction Sizes
Share of Off-chain Transactions
50%
36%
14%
2016
2017
9%
46%
46%
20%
2017
35%
45%
$0.1-$99
$100-$1,000
>$1,000
20%
46%
34%
2016
Domestic
Cross-border
SECTION 2: GLOBAL USAGE
40
Cross-border transactions are large, with approximately half of all transactions between $100 and
$1,000. Both domestic and cross-border transactions have seen a trend toward higher-value payments,
with the share of under $1,000 payments declining from 66% in 2016 to 55% in 2017.29 Whether this is
due to the increase in the price of cryptoassets, surge in fees, or greater acceptance of cryptoassets in
large transactions is beyond the scope of this report.
Consumer payment volumes remain at low levels
The study also finds that the median size of the average consumer-to-consumer (P2P/C2C) transfer
in our survey is $100, consumer-to-business (C2B) payment is $14, while the average business-to-
business transfer (B2B) is significantly larger ($50,000). These values imply that cryptoassets are used
to transfer funds "within" user types, rather than purchase goods and services. While tens of thousands
of merchants worldwide are purported to accept cryptoassets for payment, reported merchant volumes
have been relatively low, which suggests that at present cryptoassets have not managed to establish
themselves as widely used currencies and general payment method for consumer payments.30
Decentralised Applications and Timestamping
"Decentralised applications" (dApps) applications that are supposed to run in a decentralised fashion
with no central operator exploded with the release of Ethereum's ERC-20 standard.31 Ethereum
currently hosts over 2,000 dApps, with a total average of 50,000 daily active unique addresses.32 Other
dApp-focused systems such as EOS and Tezos have experienced substantial growth as well.
While dApps encourage the use of cryptoassets as payments for services or goods, their user and
transaction volumes are still very low: they are an emergent phenomenon, not substantial enough to
drive cryptoasset use. DEXes are the most popular dApp category, generating most transaction volume
and attracting most users.
Timestamping constitutes another, non-monetary application of public blockchains. Large amounts of
"off-chain" data can get anchored into networks such as Bitcoin and Ethereum, which can act as global
public notaries enabling anyone to independently verify the existence and integrity of the timestamped
data. A growing number of transaction outputs using Bitcoin's OP_RETURN field can be observed since
2015, suggesting that the use of public blockchains for timestamping has become more popular.33
Speculation and Investment
The volatile nature of cryptoasset markets make them a desirable target for speculators: available data
suggests that cryptoassets display behaviour consistent with speculative investment rather than use as
currency or payment methods. Reported global exchange volumes frequently amount to US$12 billion a
day, with over-the-counter (OTC) volumes estimated by some to be between two and three times larger.34
Nevertheless, the figures above significantly dwarf observed on-chain transaction volumes, further
29
This seems consistent with on-chain payments analysis (see Figure 19).
30
Jonkers, N. (2018) What drives bitcoin adoption by retailers? DNB Working paper 585. Available at: https://www.dnb.nl/en/news/
dnb-publications/dnb-working-papers-series/dnb-working-papers/Workingpapers2018/dnb373270.jsp [Accessed: 02 December
2018].
31
However, it is unclear whether the majority of currently active dApps can be considered "decentralised" given the presence of
centralised developer teams.
32
Data available at: https://www.stateofthedapps.com/ [Accessed: 02 December 2018]. Readers should recall that a single user may
have multiple addresses, meaning that the number of daily active addresses does not necessarily reflect the number of unique
users.
33
Data available at: https://opreturn.org/ [Accessed: 02 December 2018].
34
TABB Group (2018) Crypto Trading: Platforms Target Institutional Market. Available at: https://research.tabbgroup.com/report/v16-
013-crypto-trading-platforms-target-institutional-market/ [Accessed: 02 December 2018].
2nd Global Cryptoasset Benchmarking Study
41
supporting the view that speculation and long-term investment currently remain the major use case for
cryptoassets.
Speculation remains the major cryptoasset use case
On average, the number of available trading pairs has grown from around 6,500 to more than 9,000 in
2018 Q4, with some exchanges providing support for more than a thousand cryptoasset trading pairs.35
According to a recent report from data service provider CryptoCompare, cryptoasset-only exchanges
are responsible on average for approximately three quarter of total spot market volumes.36 Bitcoin-to-
fiat (and vice-versa) volumes in 2017 Q4 have been dominated by USD trading (roughly responsible for
half of volumes), followed by the Japanese Yen (21%) and the Korean Won (16%), with the latter showing
significant growth since October 2018. Unfortunately, limited data availability, in conjunction with
inconsistency across datasets and data collection methods, hinder conducting further data analysis of
this type of usage.37
There is a distinction to be made between investors with different time horizons: some take a short-
term perspective and actively trade cryptoassets often on a daily basis, whereas others adopt a more
long-term oriented view. While the former are primarily composed of traders engaging in short-term
speculation, the latter are long-term holders that comprise retail investors, high net worth individuals
(HNWIs), and institutional investors. These can include both new crypto-focused investment funds and
more traditional funds, as well as family offices.
Given high price fluctuations on the cryptoasset trading market, most exchanges are reluctant to offer
leverage on trades to their investors. Yet, trading on margin is made available to cryptoasset-investors by
some service providers: among surveyed participants, some exchanges provide leverage of 2x whereas
others offer up to 100x, with the average amount of leverage being 27x (median 3.3x).
Cryptoasset exchanges offer a median amount of leverage of 3.3x
However as high leverage rates are seen as an unscrupulous business tactic, financial regulators in most
countries enforce limits. The Japan Virtual Currency Exchange Association (JVCEA), a self-regulatory
organisation of the Japanese cryptocurrency industry, has proposed a cap on the leverage offered by
cryptocurrency exchanges (1:4).38
Bitcoin futures have become an integral part of the cryptoasset investment landscape since Chicago-
based exchanges CME Group and Cboe started offering cash-settled Bitcoin contracts in December
2017.39 These products allow industry actors, such as miners and payment service providers as well as
investors, to hedge the volatility inherent in cryptoasset markets and contribute to the growing maturity
of cryptoasset-related services.
35
Data available at: https://www.blockchaincenter.net/cockpit/stats/ [Accessed: 02 December 2018].
36
CryptoCompare (2018) CCCAGG Exchange Review. Available at: https://blog.bitmex.com/wp-content/uploads/2018/11/
cryptocompare_exchange_review_october_2018.pdf [Accessed: 02 December 2018].
37
Exchange volumes are frequently reported differently across data service providers, and some exchanges tolerate wash trading to
artificially increase volumes.
38
Data available in the Japanese Virtual Currency Exchange Association's (JVCEA) report presentation to the Financial Services
Agency (FSA) during the Study Group Meeting on the Virtual Currency Exchanges held in April 2018. A summary is available
at: https://www.coindesk.com/japanese-crypto-exchange-group-gets-legal-status-to-self-regulate-industry/ [Accessed: 02
December 2018].
39
Some cryptoasset exchanges have been offering derivative products a long time before this date.
SECTION 3: GATEWAYS AND ECONOMIC CONNECTIONS
42
SECTION 3: GATEWAYS AND
ECONOMIC CONNECTIONS
Connecting the Ecosystems
The cryptoasset ecosystem comprises thousands of cryptoassets that each have their own local
ecosystem. Cryptoasset gateways provide interfaces for users to seamlessly move between local
ecosystems. These entities do not handle fiat currency and provide exclusively cryptoasset services.
In contrast, fiat gateways connect the cryptoasset ecosystem with traditional markets: examples include
payment service providers that allow cryptoasset holders to purchase items at retail stores, or exchanges
that enable users to make deposits or withdrawals into traditional banking or financial systems. An on-
ramp is a fiat gateway that lets users to convert fiat currency into a cryptoasset, while an off-ramp allows
user to convert a cryptoasset into fiat currency.
Figure 19: Conceptual mapping of monetary flows between ecosystems
Figure 19 provides an approximation of the major flows of funds between the various ecosystems. Flows
can be denominated in cryptoassets or fiat currency, and do originate from various actor types.
Cryptoasset Ecosystem
Users
Storage
Providers
Multi-
segment
Payment
Service
Providers
dApps/
ICOs
Merchants
Investors
Crypto-only
Exchanges
Ethereum
Ecosystem
Bitcoin
Ecosystem
Consumers
Businesses
Miners
Miners
Exchanges
Cryptoasset users
On/off-ramps
Fiat gateways
Cryptoasset gateways
Cryptoasset flows
Fiat currency flows
2nd Global Cryptoasset Benchmarking Study
43
On-Ramps and Off-Ramps
Exchanges are the major on- and off-ramps to the cryptoasset ecosystem, with the majority of exchanges
in our sample accepting national fiat currencies for use in cryptoasset trades. Smaller exchanges are
more likely to provide cryptoasset-only exchange services than large exchanges.
Only 19% of surveyed small exchanges and 9% of large exchanges
support exclusively cryptoasset-to-cryptoasset trading
Figure 20 reports the usage rates of various fiat currency deposit and withdrawals methods for
exchanges, payment service providers, and multi-segment firms (in this context companies that blend
aspects of exchanges and payment service providers). In all cases, bank wires receive the most support,
with cash, debit, and credit cards the second tier of acceptability. While Figure 20 is aggregated data,
results are similar when examined across regions.
Figure 20: Bank wires are the most supported payment method for both deposits and withdrawals
The popularity of multiple means of deposit and withdrawal methods hide the fact that most firms
only accept a few. Across survey respondents, 36% reported accepting a single payment method for
deposit, while 51% allowed only one payment method for withdrawals. At the other end of the spectrum,
one in five firms allow four or more methods to deposit fiat currency, while only 12% allow four or
more payment methods for withdrawal. We can observe a pattern of service providers offering fewer
withdrawal than deposit methods , with the exception of multi-segment firms. This suggests that it is
currently easier for investors, users, and consumers to enter the cryptoasset ecosystem than to leave.
Entering and maintaining good banking relationships remains a major challenge
Establishing and maintaining banking relationships is of great concern to exchanges as bank wires are a
popular method of funding and withdrawing funds. Both small and large surveyed exchanges find that
the inability to enter a bank relationship poses a risk to their operations - a concern that has increased
considerably since 2017 for large exchanges. For small exchanges, their concern about entering and
maintain banking relationships is behind concerns about IT security and regulatory burden, and equivalent
to fraud and inability to find talent.40
40
See Table 6 in Appendix
Fiat Currency Deposit and Withdrawal Methods
Bankwire
Cash
Credit card
Debit card
ATM
Online transfer
Other
85%
87%
41%
37%
24%
20%
25%
22%
46%
24%
14%
17%
9%
7%
Multi-segmentBankwire
Cash
Credit card
Debit card
ATM
Online transfer
Other
83%
82%
28%
6%
44%
41%
11%
12%
28%
0%
17%
12%
6%
6%
Exchange-onlyBankwire
Cash
Credit card
Debit card
ATM
Online transfer
Other
100%
100%
0%
50%
0%
0%
20%
0%
0%
50%
0%
0%
40%
50%
Payment-only
SECTION 3: GATEWAYS AND ECONOMIC CONNECTIONS
44
Figure 21: Nearly a third of payment service providers have no existing relationships with
established payment network
Payment service providers allow cryptoassets to be directly converted into goods and services without
requiring conversion into a national fiat currency. Figure 21 shows that only 12% of surveyed payment
service providers do not have any existing relationships with established payment networks (e.g. bank
transfers, card networks, point of sale, terminals, ATMs, mobile money, eCommerce platforms), and are
currently not in active discussions to build partnerships. The remaining 88% were either actively engaged
in discussion to establish a relationship (18%), or already established a relationship (69%). Among those
with a relationship (48% of survey respondents) 70% were in an active discussion to expand their
relationship.
Relationship with Established Payment Networks
Share of Service Providers
Existing and in active
discussions to expand
48%
Not existing and not
in active discussions
12%
Not existing but in
active discussions
18%
Existing and not
planning to expand
21%
2nd Global Cryptoasset Benchmarking Study
45
Internal Cryptoasset Ecosystem Flows
Once actors have entered the cryptoasset ecosystem via on-ramps, they can start transacting with
cryptoasset-denominated funds. There are two major ways of transacting within the cryptoasset
ecosystem: users can initiate direct (on-chain) transactions that will clear and settle on the respective
blockchain, or they can do indirect (off-chain) transactions using service providers.41
Figure 22: Small exchanges have a larger relative share of outgoing transactions than large
exchanges
Figure 22 shows that only 10% of transactions initiated from users of large exchanges are directed at
external wallets such as service providers, merchants, or user wallets, against 37% for small exchanges.
All remaining transactions (63% of small exchanges and 90% of large exchanges) occur off-chain within
internal exchange recordkeeping systems.
As internal transfers between exchange accounts constitute 15-18% of all transactions for both large
and small exchange, the primary difference comes from exchange open-market (e.g. order-book)
transactions. For large exchanges, 75% of transactions are directed at the open market within the
exchange, while it is only 45% of transactions for smaller exchanges. This could suggest that larger
exchanges have a bigger percentage of passive investors who will leave funds at the exchange as opposed
to more active traders who tend to use smaller exchanges. On the other hand, it could also reflect that
larger exchanges simply have a larger number of transactions on their open market.
Only 18% of transactions from exchanges operating across multiple segments go to an external wallet,
compared to 45% for specialised exchanges. This strengthens the theory that multi-segment entities are
one-stop shops for storing and managing user funds without having to switch services where most users
stay on the platform.
This finding has public policy implications: most fiat-supporting exchanges are regulated or at the
very minimum have to comply with existing regulations. As a result, regulators have a certain level of
oversight and control over these entities; unlike cryptoasset-only exchanges, which tend to be regulated
significantly less frequently (although roughly half are engaged in self-regulation in some form and to
some extent see Section 5 for more information).
41
Layer-2 solutions such as the Lightning network enable direct off-chain transfer of cryptoassets without going through a service
provider(i.e. trust-minimised).
Transaction Destination
Share of Transactions
45%
18%
37%
Small exchanges
75%
15%
10%
Large exchanges
Open market
Internal transfers
External wallet
SECTION 3: GATEWAYS AND ECONOMIC CONNECTIONS
46
Figure 23: Exchanges report that fiat-to-cryptoasset (and vice-versa) transactions still make up the
majority of total exchange trades
Publicly available data on the popularity of fiat-to-cryptoasset trades (and vice-versa) relative to
cryptoasset-to-cryptoasset trades is contradictory across sources. Within our survey, Figure 23 shows
that both large and small exchanges reported primarily processing fiat-to-cryptoasset trades, with both
the number of transactions and their volume reflecting similar shares.
The fiat in- and out-flows of exchanges outnumber internal transaction activity, which may suggest that
exchange-based speculation remains the dominant cryptoasset use case: most users are not purchasing
cryptoassets and using them as a replacement for means of payment.
2017 Exchanges Internal Currency Mix
Share of Internal Exchange Transactions
10%
63%
27%
Number of transactions
Large Exchanges10%
62%
28%
Transaction volume ($)
Fiat <> Fiat
Fiat <> Cryptoasset
Cryptoasset <> Cryptoasset
79%
21%
Small Exchanges83%
17%
Number of transactions
Transaction volume ($)
2nd Global Cryptoasset Benchmarking Study
47
Figure 24: Payment service providers use cryptoassets less frequently for fiat-denominated cross-
border payments
The dominant use of fiat currency as a primary on- and off-ramp presents an interesting contrast to
payment service providers. Figure 24 shows that payment services reported a decrease in the use of
cryptoassets to facilitate fiat-denominated cross-border payments (i.e. using cryptocurrency as a vehicle
currency) in 2017 compared to 2016. However, the share of cryptoasset-to-cryptoasset transfers (i.e.
directly paying recipients in cryptoassets as opposed to fiat currency) has significantly increased. The
payments landscape is still dominated by fiat-to-cryptoasset transfers allowing users to buy and sell
cryptoassets with fiat currencies without having to leave the payments platform.
4% of payment service providers are using cryptoassets exclusively as
vehicle currencies to facilitate fiat-denominated cross-border transfers
Managing Volatility
Cryptoasset payment service providers may engage in varying activities, but they all share in common
that they handle cryptoassets. This can expose companies to significant volatility risk. More than half
of payment service providers report reducing volatility risk by simply buying the required cryptoasset
on-demand when necessary a "Just-In-Time" conversion strategy. The use of cryptoasset derivatives
for managing volatility risk is very rare: only a quarter of payment services require customers to put up
margin. Futures contracts are available in some locations, but primarily focus on Bitcoin.
Payment Service Providers - Internal Platform Transactions
Share of Internal Transactions
23%
60%
17%
2017
Transaction volume ($)27%
67%
6%
2016
Fiat <> Fiat
Fiat <> Cryptoasset
Cryptoasset <> Cryptoasset
23%
26%
Number of transactions26%
6%
2017
2016
51%
68%
SECTION 3: GATEWAYS AND ECONOMIC CONNECTIONS
48
Payment service providers have different strategies for managing cryptoasset volatility
What are stablecoins?
The current generation of stablecoins are digital tokens that promise stable purchasing
power or a fixed conversion rate to a specific asset or commodity. Stablecoin parity can
be maintained by promising a redeemable rate using reserves (e.g. fully fiat-collateralised
such as Tether, USDC, or the Gemini Dollar), or through algorithmic monetary policy (e.g.
MakerDAO) to stabilise price fluctuations. Stablecoins allow cryptoasset users to avoid the
high volatility in prices that characterise most existing cryptoassets, while still remaining
agile within the crypoasset space.
There is a growing desire to bring stability to the cryptoasset market through the implementation
of stablecoins. These are price-stable tokens, with a market price that is pegged to another familiar
and stable asset, like the US dollar or oil. They also help reduce the friction between the cryptoasset
and fiat currency financial systems which could potentially increase access to new kinds of assets and
opportunities. While the fundamental goal of stablecoins is to reduce the cryptoasset market volatility,
they are currently mainly used by traders to arbitrage between exchanges, or by small cryptoasset-only
exchanges that have been denied banking relationships and are unable to hold any funds in fiat.
The use of cryptoasset derivatives for managing volatility risk is very rare: only a quarter of payment
services require customers to put up margin. Futures contracts are available in some locations, but
primarily focus on Bitcoin.
2nd Global Cryptoasset Benchmarking Study
49
SECTION 4: STORAGE AND CUSTODY
SEGMENT
Cryptoasset Custody
Cryptoassets use public key cryptography, where a private key comparable to an account password
is necessary to authorise ("sign") a movement of funds stored in an address. The address, which can be
thought of as an account number, is derived from the public key that mathematically corresponds to the
private key. Software programs, commonly referred to as wallets, handle the management of these key
pairs.
If a private key is stolen, the holder of the private key can fraudulently authorise transactions that the
owner would not. In the cryptoasset industry, custody refers to the business of the secure storage of these
private keys - not storage of the assets themselves.
Custodial service providers control the users' private keys of users,
while non-custodial service providers do not
Storing private keys securely can be a cumbersome task: key management is notoriously difficult and
requires a certain level of technical proficiency, which is why it is often outsourced to third-party
"custodial" service providers. A custodial service provider is one where the service provider has been
granted possession of users' private keys. In this context, transfers of cryptoassets are generally tracked
on the service provider's balance sheet rather than being verified by the blockchain (see on-chain vs. off-
chain discussion in Section 2). In contrast, a non-custodial service provider provides wallet infrastructure
but does not hold users' private keys: users remain at all times in full control of their funds and directly
transact via the blockchain.
Many exchanges and payment service providers now provide wallets as part of their service, meaning that
they can store customer funds and thus potentially act as custodial service providers.
Table 3: We distinguish between four different custody settings
TYPE
"EXTREME" SETTING
"MODERATE" SETTING
Custodial
Service provider controls
private keys and has full
control over funds.
User has the option and can choose
whether to defer custody to a third party
or remain in full control over funds.
Non-
custodial
User controls private keys
and has exclusive control
over funds.
Both user and service provider hold one
(or several) private keys so that the service
provider cannot unilaterally move funds
without user approval.
SECTION 4: STORAGE AND CUSTODY SEGMENT
50
Figure 25 shows that the distribution of custody types is very similar between firms providing exchange
services and companies providing storage services. While the majority of small service providers do not
keep custody of user funds, approximately two-thirds of large service providers manage custodial wallets
for their customers. Firms exclusively providing storage services (69%) and exchange services (48%)
more frequently opt for non-custodial methods of storing cryptoassets, regardless of company size.
Figure 25: Large companies provide significantly more often custodial services
Despite a significant share of service providers offering self-custody options to their users, the majority
of users choose not to use it; instead opting for the convenience and peace-of-mind of custodial
solutions. While keeping control of their own funds empowers holders with additional financial
independence (e.g. funds cannot be easily seized), it also imposes significant burdens on users when it
comes to protecting and accessing their funds.
Custodial service providers take full control over user funds, which can lead to some undesirable
outcomes. For instance, internal security breaches can result in the loss or theft of all customer funds,
leaving customers often with little to no recourse with regards to recovering their funds. Interestingly,
23% and 13% of small and large companies, respectively, have implemented a system where the
company cannot unilaterally move user funds but needs the users' permission first, which somewhat
mitigates this danger.
In the event of customer funds being lost or stolen, a growing number of custodial service providers have
put a refund procedure in place: 64% of custodians have a written refund procedure, as opposed to 62%
in 2017. However, Figure 26 reveals that there are significant differences between specialised exchanges
and multi-segment companies: only one third of custodial exchange-only firms have an existing refund
procedure, compared to three quarters of multi-segment firms.
Custody Types
User has the option
User control private key(s)
Service provider controls private key(s)
User and service provider hold one (or several) key(s)
Small
65%
17%
Large
13%
63%
7%
23%
7%
4%
Exchange services
Storage services
26%
49%
23%
29%
52%
17%
2%
3%
Custodian
Custodian
Non-custodian
Non-custodian
2nd Global Cryptoasset Benchmarking Study
51
Figure 26: Two-thirds of specialised custodial exchanges do not have an existing refund procedure
Note: non-custodial service providers have been removed from the analysis.
Since custodians can move a substantial share of internal volumes off-chain using internal recordkeeping
systems, there are concerns that some may run on a fractional-reserve basis. This means that the
number of cryptoassets outstanding in their books is larger than the actual amount of cryptoassets held
in custody. In such a system, the claims consumers have on the custodian exceed the cryptoassets the
custodian has immediately available in their reserves. Unlike direct P2P trades on public blockchains, a
lack of transparency resulting from the closed nature of these internal books makes it difficult for users
to verify the solvency of the services they use.
Three out of four custodians have an
external audit of their cryptoasset reserves
In order to counter the aforementioned allegations, 76% of custodial service providers report conducting
an externally-led audit of their cryptoasset reserves in the past 12 months, which stands in stark
contrast with the relative minority of non-custodians who did so (35%). In line with previous findings,
large service providers (70%) are twice as likely to perform audits than small firms (38%). Interestingly,
71% of cryptoasset-only firms do not conduct audits of their reserves, whereas 58% of fiat-supporting
companies do.
Source Code
Storage service providers have the option to release their wallet code under an open-source license or
to keep it closed-source (i.e. proprietary). This distinction is important because open-source enhances
portability, transparency, and auditability of the software, although it could also make security flaws
easier to detect to hostile actors and may decrease the ability of the provider to monetise their service.
Conversely to their proprietary counterparts, discontinued software code bases under an open-source
license are still available and usable to external parties.
Refund Procedure
Exchange-only
33%
67%
Multi-segment
75%
25%
No refund procedure
Existing refund procedure
SECTION 4: STORAGE AND CUSTODY SEGMENT
52
Figure 27: Significant differences in terms of source code openness can be observed between
storage providers
Figure 27 reveals that small storage service providers are more inclined to have partially or fully-open
source wallet code than large ones. Similarly, custodians are also more likely to have closed-source
wallets (63%) than non-custodial service providers (20%). While two-thirds of storage-only providers
have a fully or partially open-source wallet (i.e. elements of the wallet are open-source), more than half of
multi-segment entities have closed-source wallet software (mostly in the form of accounts).
Key Storage Can Take Different Forms
Key storage can take two major forms: cold and hot storage. In cold storage, the private key is stored
offline in a cold wallet that has never been connected to the Internet and thus should not have
been easily compromised. Methods of cold storage include hardware wallets and other air-gapped,
disconnected hardware devices. Conversely, hot storage refers to keeping private keys on an online
device that is connected to some network, i.e. in hot wallets. Examples of hot wallets are web-based
wallets as well as desktop and mobile wallets running on connected machines.
Cold storage is generally considered a safer form for storing private keys, since cold wallets are less
vulnerable to network-based theft and require physical access. However, there are trade-offs involved:
cold wallets are generally more cumbersome for users to access, which leads to less flexibility and
longer waiting times. Choosing between either form thus comes down to holders' need to access funds,
frequency of trading activity, and transaction amounts (hot wallets are deemed fine for storing small
amounts, whereas larger amounts should be moved into cold storage).
Source Code
Share of Service Providers
Wallet code is fully open-source
Aspects of the wallet are open-source
Wallet code is fully closed source
17%
21%
63%
Large Firms
31%
37%
31%
Small Firms
11%
63%
38%
28%
Custodians
Non-custodians
26%
34%
2nd Global Cryptoasset Benchmarking Study
53
Figure 28: The share of funds held in cold storage has slightly decreased over 2017
Figure 28 shows that the use of funds in cold storage has slightly decreased between 2017 and 2018
Q1. The average share of cryptoasset funds kept in cold storage by multi-segment entities amount
to 83%, slightly higher than companies specialised in exchange (79%) or payment services (55%), but
smaller than for entities exclusively providing storage services (100%).
From 2017 to 2018 Q1, the figure for exchange-only providers has decreased by 8 percentage points,
which is likely a result of intensified trading activities that require exchanges to have immediate access
to cryptoasset funds. Interestingly, no significant difference can be observed between custodians and
non-custodians, which suggests that trading activities have been equally distributed across both custody
types.
Multi-Signature
An additional method for securing private keys is to use multi-signature often colloquially referred to as
multi-sig. Under a multi-signature scheme, multiple keys can be combined together so that a specific fixed
number of keys is required to sign a transaction and move funds. Holding an individual key is insufficient
to enact a transaction.
Multi-signature is a powerful tool that enables new types of ownership and custody. Users can remain in
full control of their funds while having peace of mind, because one of several keys required to move funds
is stored as a back-up by a service provider who cannot unilaterally access the funds. Similarly, ownership
of a given cryptoasset can be distributed among multiple people or entities, requiring a majority of them
to reach agreement before being able to enact transactions.
Multi-signature is also often used as a complementary tool for additional security: for instance, 87% of
custodians use multi-signature techniques as part of their cold storage system (78% for hot wallets), as
opposed to only 69% of non-custodians (68% for hot wallets). The average custodian secures a lower
proportion of hot wallets (73%) via multi-sig than the average non-custodian (98%), again likely because
they need to have quicker and easier access to wallets for client disbursements.
Share of Cryptoasset Funds
2017
Average
2018 Q1
Average
87%
82%
Cold storage
Hot storage
18%
13%
SECTION 5: REGULATIONS AND COMPLIANCE
54
SECTION 5: REGULATIONS AND
COMPLIANCE
The rise and the subsequent plunge in cryptoasset prices triggered a wave of complaints from retail
investors who lost money, which in turn has prompted financial watchdogs to further investigate the
cryptoasset industry.
The regulatory landscape surrounding cryptoassets is diverse and ever-changing: some jurisdictions
have adopted a wait-and-see approach (e.g. UK, Canada) whereas other have taken either a more
proactive stance (e.g. Japan, Malta) or unsupportive position (e.g. China, India). The cryptoasset
regulatory landscape will be closely examined in an upcoming report by the CCAF.
5.1 The Impact of Regulations
There is a perception that the cryptoasset industry flaunts regulations and conducts business without
regard for any legal directives. This section will show that regulations do impact both the users service
providers choose to accept and the countries where service providers choose to do business, indicating
that this impression is at least partially incorrect.
User Impact of Regulations
Many respondents report refusing to serve customers located in a particular jurisdiction as a result
of changes in the region's regulatory environment. This is particularly high for payment-only (80%),
exchange-only (70%), and multi-segment service providers (82%). Small firms (67%) are less likely than
large firms (83%) to decline customers from a specific jurisdiction as a result of changes to the regulatory
environment with no substantial difference across regions.
Changes in the regulatory environment have had a significant impact
on the decision not to serve customers from a specific jurisdiction
Custodial firms and fiat-supporting service providers most resemble traditional financial entities: 79%
and 88%, respectively, refused to serve customers residing in particular jurisdictions. Surprisingly,
68% of non-custodians and 44% of cryptoasset-only firms indicate having had to decline customers
from specific jurisdictions after changes in the regulatory environment. These figures are higher than
expected given that these firms exist in a regulatory grey zone in many regions due to the particular
nature of their services (i.e. no custody of user funds, no handling of fiat currency).
2nd Global Cryptoasset Benchmarking Study
55
Industry Reactions to Regulations
Survey responses suggest that the regulatory environment, which includes existing regulations,
guidelines and planned changes, has a direct influence on cryptoasset businesses' decision to enter or
expand to a particular jurisdiction or alternatively to close operations and leave a given country. In
many cases, regulatory changes involve introducing regulatory guidance where none existed or were
enforced previously. Figure 29 illustrates the impact of changes in the regulatory environment on the
decision-making process of industry actors.
Figure 29: Regulatory interventions have a significant impact on operations
Note: location can refer to offices, facilities, or similar types of properties.
52% of surveyed exchange-only entities and 67% of multi-segment service providers stated that changes
perceived or actual in the regulatory environment have led them to open offices or facilities in a
new location. For instance, the ban on fiat-to-cryptoasset (and vice-versa) trading imposed on domestic
exchanges by the Chinese government in September 2017 spurred the major Chinese companies to
open offices in other jurisdictions with "friendlier" regulatory frameworks, such as Singapore or Malta.
However, there are significant differences between small and large firms: 73% of surveyed large entities
opened facilities in specific locations following a change in the regulatory environment, compared to only
44% of small entities opting for a similar strategy.
Share of Service Providers
Location Opening
Not relevant to our operations
Not relevant to our operations
No location opening
No location closure
Location opening
Location closure
Location Closure
Cryptoasset-only
Cryptoassets & Fiat
60%
28%
12%
38%
50%
12%
36%
18%
45%
11%
47%
42%
SECTION 5: REGULATIONS AND COMPLIANCE
56
Differences also lie between cryptoasset-only and fiat-supporting entities, where nearly half of
cryptoasset-only service providers indicate that the regulatory environment is not a relevant factor
in the decision-making process for opening a new location often because they contend that the
regulations do not apply to them. Companies registered in the Middle East and Africa as well as South
America have more frequently opened new locations following regulatory changes. Overall, current
changes in the regulatory environment are more likely to impact decision-making in encouraging location
opening than location closure.
The regulatory environment is also a determining factor for mining organisations when deciding whether
to open new facilities or offices in specific locations. About a third of large miners, small miners, and
individuals agreed that their immediate regulatory environment is confusing and inconsistent. Half of
surveyed large miners perceive regulation as adequate and appropriate, a statement to which only 9% of
small mining organisations and individuals agreed. Instead, 23% of small miners believe that no bespoke
regulation exists and that none is need, while 16% think that regulation is non-existent but needed.
Interestingly, no clear pattern in terms of regional differences could be observed: it is thus unclear
whether there are specific regulations that miners are reacting to in their sentiment responses.
Cryptoasset Firms Collaborate Directly with Regulators
Contrary to popular media narratives, most survey respondents indicate high levels of interaction
and collaboration with public-sector stakeholders such as regulators, policymakers, legislators as
well as standard-setting bodies with regards to cryptoasset-related regulations. For example, public
consultations and other forms of public hearings have been held between cryptoasset entities and
respective authorities in the USA, Canada, Bermuda, and Malta, among many others.42 Regulatory
innovation initiatives such as regulatory sandboxes and innovation offices have facilitated regulators'
engagement with industry actors. This type of engagement is a helpful factor for reducing the knowledge
gap and addressing regulatory issues from an industry perspective, which might contribute to more
regulatory clarity in the future.
High levels of interaction and collaboration are observed between industry and the public sector
The study finds that multi-segment entities and exchange-only firms are most likely to engage with
regulators directly, whereas less than half of storage-only companies tend to do so. 91% of surveyed
custodians and 85% of fiat-supporting entities collaborate with regulators in some capacity, as opposed
to 65% of non-custodians and 50% of cryptoasset-only companies. Similarly, large firms (89%) are more
likely to engage with regulators than small firms (67%). Geographically speaking, industry-innovator
collaboration appears to be less prevalent in Asia-Pacific than in other regions at the time of the survey.
42
There are also efforts to formalise self-regulation across the industry. Examples include the Japanese Virtual Currency Exchange
Association (JVCEA), and the Global Digital Finance (GDF) initiative.
2nd Global Cryptoasset Benchmarking Study
57
5.2 KYC/AML Policies
Enforcement of KYC/AML regulations in the cryptoasset context is a contentious issue amongst some
in the industry. Some argue that the original intention of cryptoassets was the freedom to move funds
around the globe without involvement or approval from any authority. Others opine that service
providers are not a native element of the cryptoasset framework, and that if they facilitate criminal
activities they should be prosecuted just like all other business would be.
What Is KYC/AML?
Know Your Customer (KYC) refers to due diligence activities that financial institutions
and other regulated companies must perform to ascertain relevant information from their
clients for the purpose of doing business with them. Anti-Money Laundering (AML) refers
to laws or regulations designed to stop the practice of generating income through illegal
actions. Regulated financial services providers are responsible for the implementation of
internal KYC and AML policies.

Figure 30: Half of all cryptoasset-only entities perform KYC/AML checks in some capacity
Figure 30 indicates that while nearly all surveyed fiat-supporting companies perform KYC and AML
checks (91%), half of cryptoasset-only firms do as well (52%). This is a surprisingly large share considering
that, depending on the jurisdiction and the precise type of activity, many were supposedly not bound
by KYC/AML regulations at the time of the survey. This may serve as an argument for proactive "good
practices" compliance from industry players, despite the absence of regulatory guidance and clarity.
Implementation
Although most entities implemented KYC/AML checks when they started their activities, a few began
KYC/AML verification at a later date (Figure 31). KYC/AML checks have been implemented by some
cryptoasset entities as early as 2011, and the pattern of performing KYC/AML checks continued from
2013 onwards without a clear watershed moment.
Know-Your-Customer / Anti-Money Laundering
Checks
No checks
Cryptoasset-only
Cryptoassets & Fiat
48%
91%
9%
52%
SECTION 5: REGULATIONS AND COMPLIANCE
58
Figure 31: The majority of entities immediately implement KYC/AML checks when launching
Note: entities that do not perform KYC/AML checks have been removed from the following analysis.
The majority of entities that support both cryptoasset and fiat currency use third-party support for
KYC/AML; thereby relying primarily on traditional service providers (62%) rather than blockchain
analytics companies. Overall, only one third of surveyed service providers reported using the services of
blockchain analytics companies that provide on-chain forensics of blockchain transactions.
Criteria
The share of service providers conducting KYC/AML verification for all accounts (i.e. both cryptoasset
and fiat currency accounts) is relatively consistent throughout industry segments at around 80% on
average. 91% of large entities supporting fiat currencies conduct KYC/AML checks on all accounts,
as opposed to 74% of small entities. No such difference emerges between small and large entities
exclusively supporting cryptoassets.
The vast majority of entities conducting KYC/AML checks inspect every account.
Respondents that do not check all accounts but use "other criteria" to perform KYC/AML verification
most frequently use account size or account activity as criteria for triggering additional checks. A slight
difference can be observed between fiat currency and cryptoasset accounts: service providers tend to
more often use account size as a criterion for the former (63%) than for the latter (56%).
Account size and activity are the two most popular criteria if
KYC/AML checks are not applied to all accounts.
KYC/AML Implementation
2010
Cryptoasset services launch date
2011
2012
2013
2014
2015
2016
2017
2018
2010
2013
2016
2011
2014
2017
2012
2015
2018
2019
KYC/AML implementation dateImplementation lag2019
x=y
2nd Global Cryptoasset Benchmarking Study
59
Account Suspensions and Closures
In an effort to prevent fraud and financial crimes, service providers may decide to close a customer
account (often referred to as "de-risking") or not to onboard new users after conducting due diligence.
An organisation can decide to do so to mitigate risks associated with higher-risk customers or for users
conducting suspicious activities plausibly related to money laundering or terrorism financing.
In general, large entities are more likely to discontinue or decide not to initiate customer relationship
following KYC/AML checks: all large entities that perform KYC/AML checks have had to close accounts/
refuse to open new accounts, whereas one third of small companies report no account closures.
An average of 14% of KYC checks at multi-segment entities resulted in
account closure and/or refusal to open account in the past 12 months
Entities engaged in multiple segments have more frequently decided to close an account or refused
to open an account, followed by payment-only service providers. The median percentage of account
closure/refusal following customer due diligence is 10% for multi-segment entities and 5% for payment-
only companies.
Some fiat-supporting entities have to close up to 85% of
accounts (or refuse new account openings) following KYC/AML checks
The median percentage of account closure/refusal that followed KYC/AML checks is higher among
large companies (10%) and fiat-supporting entities (9%) than small (3%) and cryptoasset-only (0%) firms,
though this range varies significantly from 0% up to 85%. This raises doubts about the quality and nature
of implemented KYC/AML programmes, which seem to vary a lot from one provider to another.
Among surveyed entities, four have account closure/refusal rates between 50% and 85%. Interestingly,
one in five respondents both cryptoasset-only and fiat-supporting entities report having 20% or more
of their KYC/AML checks resulting in them not opening a new account or closing an existing one (Figure
32).
Figure 32: 83% of fiat-supporting entities have had to close / refuse to open accounts after KYC/
AML checks
Note: entities that do not perform KYC/AML checks have been removed from the analysis.
Share of Service Providers
Cryptoasset-only
Cryptoassets & Fiat
17%
1 - 5%
6 - 10%
11 - 20%
20%+
No closure
Share of KYC/AML checks resulting in account closure/opening refusal
17%
29%
20%
17%
60%
20%
20%
SECTION 5: REGULATIONS AND COMPLIANCE
60
5.3 Compliance Team
KYC and AML obligations have dominated the debate on compliance issues facing cryptoasset
companies. However, compliance entails a broader set of rules prescribing standards in relation to,
among others, information disclosure to clients, financial statements, cybersecurity, and other prudential
aspects. Regulatory requirements vary in respect to different business activities.
While regulators are still studying the cryptoasset industry, many entities involved in the ecosystem have
taken a proactive approach to compliance. This often involves building up a dedicated compliance team to
monitor the immediate regulatory environment and help executives navigate the complexities of financial
regulations.
Figure 33 demonstrates that the use of compliance teams varies with the type of supported assets. As
expected, the large majority of entities supporting both fiat currency and cryptoassets have an in-house
compliance team, whereas only 5% outsource compliance to a third party. Whilst more than half of
cryptoasset-only firms do not have a compliance team in line with the absence of bespoke regulations
in most jurisdictions, a remarkable share (37%) have set up an internal compliance team. This move is
primarily intended to better respond to potential newly-introduced regulations specific to cryptoasset
businesses or to anticipate future regulatory changes.
Figure 33: Nearly half of cryptoasset-only service providers have a compliance team
With regards to the costs associated with compliance, nearly half of cryptoasset-only firms report
allocating a part of their budget to compliance (between 1-10% of total budget) compared to 78% of
fiat-supporting entities (with an average of 11-15% of total budget). Noticeably, 4% of fiat-supporting
companies indicated having no compliance headcount and cost (the remaining 18% did not provide any
figures).
A cross-segment comparison reveals that most service providers have an internal compliance team (71%
and 75% of exchange-only and payment-only entities, respectively, as well as 88% of entities involved
in more than one industry segment). Although a substantial majority of companies exclusively providing
storage services stated not being subject to compliance (79% selected "Not applicable"), 7% of them
indicated having an internal team dedicated to compliance.
A growing number of cryptoasset-only firms have positive compliance headcount and cost
Interestingly, some exchanges who responded "Not applicable" were, contrary to their response, subject
to existing regulations (primarily KYC and AML laws), which casts doubt upon their regulatory awareness
and raises concerns about the quality of compliance checks.
Compliance Team
Share of Service Providers
Not applicable
Other
Outsourced
In-house
Cryptoassets & Fiat
7%
1%
5%
86%
Cryptoasset-only
53%
11%
37%
2nd Global Cryptoasset Benchmarking Study
61
No significant difference was observed between large and small companies in terms of headcount; with
both having 6-10% of their employees dealing with compliance. However, compliance costs seem to be of
a greater burden for large companies: 11-15% of their budget is allocated to compliance, versus 6-10% of
small companies' budget.43
5.4 Licensing
In most jurisdictions, conducting any type of financial activity is conditional upon obtaining an operating
license. Licensing landscapes are extremely diverse, and so are the costs and the administrative burden
associated with the application process.
A limited number of jurisdictions have created a licensing regime specific to business engaged in
cryptoasset activities. For instance, cryptoasset exchanges in Japan are required to apply for the Virtual
Currency Exchange License, while those registered in Bermuda must obtain a Digital Asset Business License
from the Bermuda Monetary Authority. A single jurisdiction may require cryptoasset companies to
obtain multiple licenses to be fully compliant. The USA, where federal states have different types of
licenses (e.g. BitLicense, Money Transmitter License), is a classic example.
5% of surveyed cryptoasset-only service providers hold a license,
compared to 39% of fiat-supporting entities
Fiat-supporting entities are most likely than cryptoasset-only firms to currently hold a license, have an
outstanding application or have the intention to apply for one in the near future (Figure 34). A small
proportion of fiat-supporting entities still consider that no license is needed in the markets they operate
in or are expecting regulators to develop a bespoke regulatory framework (i.e. "N/A" respondents).
Figure 34: Most fiat-supporting entities are in the process of applying for a license or intending to
apply in the future
43
All numbers presented in this section are median figures.
Current licensing
Outstanding application
Future application
Cryptoasset-only
32%
5%
64%
4%
39%
57%
35%
30%
35%
N/A
No license
N/A
No outstanding application
N/A
No intention
Cryptoassets & Fiat
License held
Outstanding appliction
Intention to apply
38%
19%
43%
22%
39%
39%
7%
79%
13%
SECTION 5: REGULATIONS AND COMPLIANCE
62
When further segmenting the data, it appears that none of the cryptoasset-only companies who
currently operate without a license have a pending application. Nevertheless, 29% of them have the
intention to apply for a license or to register with local authorities in the near future, highlighting
their anticipation of changes in the regulation and the current regulatory uncertainty they are facing.
In contrast, 27% of fiat-supporting entities that currently not holding a license have an outstanding
application and 88% are planning to file a license application in the future.
Licensing-related figures greatly vary across industry segments. Entities active in multiple segments are
more likely to hold a license (44%) or to have applied for a license (36%). Noticeably, none of the surveyed
storage-only providers currently hold a license, which can be explained by the fact that all surveyed
wallet providers are exclusively handling cryptoassets. Nevertheless, one third contemplate filing for
a license in the future; which is the case for the majority of exchange-only, payment-only, and multi-
segment service providers.
Large firms (45%) are twice as likely as small firms (21%) to currently hold a license, although no major
differences can be observed in terms of future applications. Furthermore, 49% of custodians currently
operate under a license, 90% of which do not intend to apply for one in the future. This greatly contrasts
with the percentage of non-custodial service providers holding a license (17%) or planning to file an
application (74%).
2nd Global Cryptoasset Benchmarking Study
63
SECTION 6: IT SECURITY
Stolen Funds Remain a Major Issue
IT security remains a major risk factor for all cryptoasset actors. Cryptoasset holders are obvious targets
for hackers and criminals as blockchain transactions are generally irreversible once the cryptoasset
leaves the intermediary (e.g. exchange or storage provider).
According to Figure 35, exchange and storage service providers alone have accounted for the loss of
more than $1.5 billion of cryptoasset funds as a result of 58 identified security breaches. The actual
number of hacks and security breaches - and the amount of lost and stolen funds would be significantly
higher if exit scams, the exploit of vulnerabilities in smart contracts, and unreported service provider
hacks were to be included.
Figure 35: More than $1.5 billion worth of cryptoasset funds have been lost as a result of security
breaches at exchanges and storage providers
Note: data sourced from a combination of publicly available sources including Rados, CoinTelegraph, and CoinIQ. Exploited
vulnerabilities in smart contracts (e.g. The DAO, Bancor) and exit scams have not been included.
Consequently, survey data indicates that concerns over cyber security risks have increased since the
beginning of 2017. Large exchanges seem to be disproportionately concerned, presumably because
rising prices and increasing amounts of funds in custody have made them lucrative targets for criminals.
Challenges for securing internal systems and cryptoasset holdings are not limited to exchanges, though:
miners also report growing concern over IT security risks, significantly up from last year.
Security techniques and methodologies are important determinants for traditional third parties such
as banks and other financial institutions when deciding whether to enter new or maintain current
business relationships.44 To address IT security concerns, the cryptoasset industry and wider community
44
An example constitutes Nonghyup Bank's reported refusal to renew its relationship with local South Korean exchange Bithumb
following the exchange's security breach in June 2018. See Young-sil, Y. (2018) Bithumb to Stop Using Real-name Virtual Accounts
to Customers from August. Business Korea. Available at: http://www.businesskorea.co.kr/news/articleView.html?idxno=24028
[Accessed: 03 December 2018].
Identified Security Breaches Leading to Loss of Funds
Exchanges and Storage Providers (excl. dApps)
$300
$200
$100
$500
4
0
$600
2
$700
8
$800
10
$900
12
$400
$0
Cryptoasset funds lostNumber of security breaches per year2011
$11m
2015
$8m
2012
$1m
4
5
6
12
6
5
10
2016
$81m
2014
$485m
2018
$854m
2013
$9m
2017
$113m
6
10
Total funds lost (lhs)
Number of identified breaches (rhs)
Million
SECTION 6: IT SECURITY
64
has tackled standardisation of cyber security practices as early as 2014 by issuing Cryptocurrency
Security Standards (CSS) to complement existing standards such as ISO 27001.
IT Security Team
On average, large entities have a smaller share (6-10%) of their full-time staff being IT security
professionals than small entities (11-20%), which may simply be a result of having a substantially larger
workforce in general that distorts relative comparisons. However, the cost structure between large and
small firms is very similar, with on average 11-20% of total budget allocated to IT security.
Differences in IT security headcount but similar budget
shares are observed between small and large firms
Interestingly, significant differences arise when comparing industry segments: entities exclusively
providing exchange and wallet services employ a larger share of IT professionals than payment-only
service providers. Similarly, they also spend more on IT security: the majority of exchange-only and
storage-only firms allocate more than 11% of their budget to IT security, whereas only 25% of payment-
only do so. On average, entities engaged in multiple segments have 11- 20% of their staff working on
cyber security; a similar share of their budget is dedicated to IT security.
Entities exclusively providing storage services have the
highest headcount and budget share dedicated to IT security
As a preventive measure against potential future security breaches, companies have established
dedicated training programmes, i.e. educational programmes for staff to understand major security
risks (e.g. social engineering attacks, password security, general security principles) and other attack
vectors. These programmes can be part of the employee onboarding process, but are often conducted
on a regular basis. As pointed out by one survey participant, training programmes are a mandatory
specification of international information security standards (ISO 27001).
Figure 36: Training programmes for staff are a common industry standard
Given that the majority of security breaches can be attributed to employee negligence and/or
wrongdoing, it is unsurprising to see that the provision of regular staff training programmes has
become a common industry standard across all industry segments (Figure 36). Companies exclusively
providing storage services clearly stand out: all surveyed entities indicated running training programmes.
Training Programme for Staff
Multi-segment
18%
82%
Exchange-only
11%
89%
No training programme
Running training programmes
Storage-only
100%
Payment-only
67%
33%
2nd Global Cryptoasset Benchmarking Study
65
Remarkably, no striking difference can be observed between custodial and non-custodial service
providers: around 80% of service providers within both categories have training programmes. As
expected, large firms are more likely to run training programmes than small ones, though to a limited
extent.
Frequency of training programmes do not only vary from one service provider to another, but also
within a given company from one employee to another (e.g. technical and non-technical staff). One
surveyed company noted that "basic security training for all non-technical staff happens every 3 months.
Advanced training for tech staff is whenever possible."
Support for staff training programmes has increased from 79% in 2016 to 89% in 2017
Security Audits
Security audits generally cover three dimensions of a company's activities: people, processes, and
technology. After conducting a risk-based assessment, security auditors collect evidence on a wide-range
of cyber security aspects, from risk management to human resources security and cryptographic controls.
Audits can be either performed internally or externally. External audits are undertaken by an independent
third party to identify security loopholes and vulnerabilities that may have been overlooked during
internal security audits. In some jurisdictions (e.g. Japan), on-site inspections and security auditing
processes by a third party, alongside a more general audit, are mandatory to apply for an operating license.
There is a general reluctance across the industry to talk about security-related issues
A significant number of respondents did not answer the question about the frequency of security audits.
With regards to external security audits, more than half of exchange-only and payment-only entities
responded "Not applicable" or "Decline to respond". The figures are somewhat lower for storage-
only service providers (35%) and entities engaged in multiple segments (37%). Respondents are fairly
more willing to share information about internal security audits, but absence of responses remains
considerably high: 44% for exchange-only service providers, 39% for storage-only service providers,
54% for payment-only service providers, and 28% for firms engaged in multiple segments.
This finding either conveys companies' reluctance to disclose information about their security auditing
processes or a lack of awareness about security audits from the surveyed representative. A critical
reader would interpret companies' reluctance as signalling the absence of any formal security verification
standards. While disclosure would have shed light on best practices, absence of information reinforces
common belief that some entities in the space do not follow security best practices.
SECTION 6: IT SECURITY
66
Figure 37: Internal security audits are conducted relatively frequently
Unsurprisingly, internal audits are conducted on a more regular basis than external audits (Figure 37).
Multi-segment and exchange-only firms conduct both external and internal security audits on a more
regular basis than other firms. However, storage-only entities are closely following when it comes to
internal audits, while payment-only firms are lagging behind.
Nearly half of small companies perform an external audit on an annual basis, as opposed to 29% of large
firms. However, small firms that conduct external audits several times a year do it more regularly than
large firms: 13% of small firms indicate undertaking an external audit on a weekly basis, compared to only
4% of large firms.
Figure 38: The public sharing of information about audits is not common industry practice
As foreshadowed by the low response rate to the security audit question, entities very rarely share
information with the public about their security audit results (Figure 38). Entities that do share
information publicly are more likely to report on externally-led audits (e.g. announce that an audit has
taken place). Protection against potential attackers by disclosing as little information as possible is one
possible interpretation.
Weekly or more often
Bi-Annually
Monthly
Annually
Quarterly
Internal audit
External audit
60%
20%
40%
20%
20%
40%
Payment-only
45%
22%
22%
10%
50%
10%
30%
11%
Storage-only
36%
5%
34%
25%
Multi-segment
Exchange-only
29%
10%
4%
22%
31%
35%
8%
20%
25%
12%
29%
24%
35%
8%
8%
Share of Service Providers
Audit Information Sharing with Public
No sharing
All audits
Externally-led audits only
Internal audits only
Storage-only
70%
30%
Payment-only
83%
17%
Exchange-only
92%
8%
Multi-segment
10%
80%
6%
4%
2nd Global Cryptoasset Benchmarking Study
67
Only a limited number of payment-only and storage-only service providers (17% and 30%, respectively)
share information about externally-led audits. Entities specialising in the exchange segment appear to
share the least information of all service providers: only 8% publish information on security audits (both
internal and external).
The likelihood of information sharing only slightly varies with company size: 25% of large firms indicate
disclosing information about security audits, compared to 16% of small firms. A comparison with 2017
survey data suggests that there is a trend to share less information: while one third of large exchanges
reported in 2017 to publish data on security audits, none of surveyed large exchanges did so in 2018.
Internal Policies
Figure 39 outlines that a majority of service providers have an internal policy on both access to sensitive
customer information (e.g. identification documents, bank details) and access to production environment
(e.g. private keys).
Figure 39: Cryptoasset companies are more likely to have written policies regarding access to
customer-sensitive information than production access
In line with previous findings, payment-only firms are less likely to have written policies on both sensitive
information access and production access than companies operating in other segments. Figures
for custodians and non-custodians are mostly similar, as both categories report having a policy for
production access in place. Interestingly, 13% of custodial service providers do not have a written policy
on sensitive customer information access, while all non-custodians do.
Internal Policies
Sensitive Information AccessProduction AccessMulti-segment
5%
95%
16%
84%
No policy
Existing policy
Storage-only
100%
100%
Payment-only
83%
17%
60%
40%
Exchange-only
25%
75%
5%
95%
SECTION 7: MINING SEGMENT
68
SECTION 7: MINING SEGMENT
7.1 The Power of Miners
Miners are the entities that are involved in the processing of transactions on public blockchains by
deciding which transactions will be added often in a single batch called a "block" to the global ledger
("blockchain"). This process generally requires attaching a financial cost to each miner's vote on the next
block in order to prevent Sybil attacks.45
In proof-of-work (PoW), the first and most common Sybil prevention mechanism used by cryptoassets,
the costs come in the form of special equipment and electricity required for solving cryptographic
puzzles. In proof-of-stake (PoS), an alternative mechanism recently rising to popularity but still mostly
experimental, the costs are modelled in the form of a "token deposit" provided by miners that can be
destroyed ("slashed") in case misbehaviour and fraud are detected.
PoW posits that miners have a vote proportional to the computing power (hashpower) they provide,
whereas PoS gives miners a vote proportional to the "stake" they provide as deposit. At the time of
writing, PoW remains the dominant mechanism used by most cryptoassets. The remainder of this section
will thus cover primarily PoW miners rather than PoS stakers.
Cryptoasset Selection
Access to public blockchains and broader DLT systems is unrestricted and permissionless. This means
that miners can decide to enter any cryptoasset system and participate in transaction processing. Survey
data shows that the majority of miners are mining more than a single cryptoasset, although significant
differences can be observed between small and large miners (Figure 40).
Figure 40: Large miners have a significantly greater diversification than small miners in terms of the
number of cryptoassets they mine
45
Sybil attacks refer to an entity creating multiple fake identities in order to rig a vote.
Number of Cryptoassets Mined
1 cryptoasset
2 cryptoassets
6+ cryptoassets
3 cryptoassets
4-5 cryptoassets
Small Miners
20%
5%
45%
30%
Large Miners
20%
60%
10%
10%
2nd Global Cryptoasset Benchmarking Study
69
Large miners tend to mine a greater number of cryptoassets: 60% of surveyed large companies mine
six cryptoassets or more, whereas no small-scale miner is operating across more than five cryptoasset
systems. This suggests that scale in the mining segment is an effective factor for reducing barriers to
entry to local cryptoasset ecosystems.
When it comes to determining which cryptoassets to mine, both small and large miners indicate that the
price is the most important criterion although large miners also consider market capitalisation, daily
reward amount, and cryptoasset price as equally important (Table 4).
Table 4: Large miners consider a larger set of criteria to determine what cryptoassets to mine
Somewhat surprisingly, the energy requirement for mining a particular cryptoasset (and therefore the
associated costs) appear to be a less important decision factor even more so for small miners. Three
quarters of large miners also take into account the reputation of a cryptoasset project before engaging in
mining. Low levels of competition are preferred to high competition, as these cryptoassets may provide
an opportunity for wondrous sudden surges in price (and therefore profitability).
Interestingly, nearly one in five small miners mine coins out of personal affection or because they
are ideologically inclined to a coin's philosophy. This seems to be in line with the finding that miners
generally tend to be "loyal" to the cryptoasset(s) they mine: only 13% and 7% of large and small miners,
respectively, have discontinued mining a specific cryptoasset since January 2017.46 This suggests that
most miners are not constantly switching between different coins.
A majority of miners continue mining the same cryptoasset(s) since early 2017
Miners use a variety of communication channels in order to stay up-to-date with the latest developments
in the local cryptoasset ecosystems that they support (Figure 41). Large miners use significantly more
communication channels than small miners and individuals: while the latter primarily rely on press
releases and direct communications among themselves to keep abreast of recent developments, large
miners also gather significantly more intelligence from conferences, social media (e.g. Twitter) and
investors directly.
46
The reasons for stopping mining a particular coin are manifold but can be grouped into the following three categories: rising costs
(and/or falling prices) eating away profit margin, change to PoW algorithm, and coin death.
Selection Criteria for Choosing Cryptoasset to Mine
Large Miners
Small Miners
Market capitalisation
88%
38%
Daily reward amount
88%
52%
Price of cryptoasset
88%
76%
Reputation
75%
29%
Energy requirement
63%
24%
Proof system
63%
29%
Low number of other miners/mining pools
25%
19%
Large number of other miners/mining pools
13%
10%
Ideology/personal affection
13%
19%
Friends/colleagues recommendation
13%
5%
SECTION 7: MINING SEGMENT
70
Figure 41: Miners use a variety of off-chain communication channels to stay updated with regards
to the latest developments of the cryptoassets they mine
Influence on Decision-Making Process
Concerns have emerged over time in the ecosystem over the role of miners and their relative influence
on the decision-making process of the DLT systems they support. These concerns were illustrated in the
August 2017 split of Bitcoin Cash (BCH) from the Bitcoin (BTC) network over disagreements on whether
users or miners should be in control of protocol governance.47
Recent forking episodes have demonstrated the multi-dimensional aspect of public blockchain
governance, which involves different groups of entities (e.g. exchanges, wallets, miners, developers,
users, merchants) keeping a check on each other. As a result, it appears that miners have become more
divided in 2018 with regards to how they perceive their influence on protocol governance (Figure 42).
47
In the meantime, Bitcoin Cash itself forked into Bitcoin ABC (BAB) and Bitcoin SV (BSV) in mid-November 2018. At the time of
writing, it appears that Bitcoin ABC has won the "hash war" and is recognised as the successor of the original Bitcoin Cash system
by receiving the original BCH exchange ticker.
Communication Channels
Conferences
Twitter
Other Miners
Large
Large
Large
Small
Small
Small
Individuals
Individuals
Individuals
50%
57%
82%
44%
43%
82%
11%
14%
82%
Internet Forums
Press Releases
Investors
Other Individuals
Involved in the Project
Large
Large
Large
Large
Small
Small
Small
Small
Individuals
Individuals
Individuals
Individuals
73%
43%
33%
6%
64%
67%
61%
36%
5%
6%
10%
64%
2nd Global Cryptoasset Benchmarking Study
71
Figure 42: Miners are becoming more divided on their impact on protocol governance
While 70% of large miners believed that their influence on protocol governance was either high or very
high in 2017 before the Bitcoin Cash fork only 46% of large-scale miners do so in 2018. This change
in figures mirrors their growing concern about unexpected changes to the protocols of the cryptoassets
they are mining (e.g. a PoW change that would make their equipment worthless), an issue that has
become more relevant given the increased frequency of hard forks since mid-2017.
In contrast, the share of small miners believing they have a very high influence on protocol governance
has grown since 2017, exceeding that of large miners. Nevertheless, unlike last year's survey no major
differences can be observed between small and large miners. Even though the majority of both small and
large miners feel that they have at least some influence on protocol governance (i.e. "medium" or above),
overall miners appear less confident in their ability to shape protocol governance than they did in 2017.
Concentration Concerns
Another major concern in the ecosystem is the (perceived) growing concentration of mining in the hands
of a few entities (Table 9 in Appendix). The fear of increased mining centralisation goes against the main
objective of cryptoasset systems; notably the ability to remain free from third-party control.
Miner Influence on Protocol Governance
Very low
Low
Medium
High
Very High
10%
20%
30%
40%
2017
23%
2018
15%
31%
23%
8%
Large Miners
3%
5%
41%
35%
16%
2017
Small Miners
(incl. Individuals)
2018
18%
27%
25%
27%
3%
SECTION 7: MINING SEGMENT
72
While these concerns are certainly justified, they often tend to myopically focus on "miners" as a whole
without distinguishing between different activities. Contrary to popular beliefs, miners are not a
homogenous group of entities that all perform the same tasks. Instead, miners can engage in a variety of
activities across the mining value chain (see Figure 3); activities that can be substantially different one
from another.
Three major types of mining concentration need to be taken into consideration
As a result, it is important to separately analyse each of the activities that are relevant to mining
concentration and identify the actors that occupy a dominant position in these respective areas. In
essence, concerns about concentration risks in cryptoasset mining can be grouped into three main
categories:
1. Hardware manufacturing concentration
The market for the manufacturing of specialised mining hardware equipment that will be used
to solve the cryptographic PoW puzzles. The analysis requires considering both the number of
competing hardware manufacturers as well as their geographic location.
2. Hashing facility concentration
To determine whether hashing is concentrated, the ownership of hashing facilities (also often
referred to as mining farms), the entities owning and/or operating these facilities (called hashers),
and their geographic distribution need to be inspected.48
3. Pool concentration
Hashers can contribute hashpower to mining pools.49 Pool operators decide which transactions to
include (or not to) in a new block before sending the candidate block to connected hashers. Both
pool structures and their geographic distribution need to be investigated.
The following subsections will cover each of the three mining concentration types by taking a deeper
look into the respective entities engaged in the related activities.
7.2 Hardware Manufacturing
Solving a cryptographic PoW puzzle amounts to finding a hash whose value remains below a specific
target value.50 This can be compared to a lottery, where there is one winning ticket that gets randomly
chosen: the more tickets you have, the more likely you are to win the lottery. Similarly, the more hashes
you produce, the more likely you are to find a valid solution to the PoW puzzle and receive the block
reward.
48
Hashing facilities are akin to data centres specialised for cryptoasset mining: they can host thousands of mining machines that
generate hashpower.
49
Hashers connect to mining pools and contribute hashpower in order to smoothen pay-out rates, as pooled mining increases the
likelihood of finding a new block and thus receiving the block reward.
50
A hash is a bit string of fixed size that is generated by running a certain input of any size through a cryptographic hashing function.
Changing a single bit of the input will result in an entirely different, unpredictable hash (i.e. output). Cryptographic hashing
functions are often use to quickly prove data integrity.
2nd Global Cryptoasset Benchmarking Study
73
Mining Equipment and Algorithms
Hashing is the process of using machines to generate hashes as a potential solution to a PoW puzzle.
Hashers are miners who own and operate machines that generate hashes. These machines can range
from simple general-purpose computers to application-specific ASICs that are optimised for performing
one single task.
General-purpose machines are devices that can perform multiple tasks and can thus be repurposed for
other applications. Application-specific machines are devices that are specifically designed to perform a
single task very well; thus, they cannot be repurposed and used for other applications (e.g. mining other
cryptoassets that use a different algorithm).
Different Hardware Types
Different types of hardware equipment are being used for cryptoasset mining. They can
generally be grouped into the following four categories:
1. CPU: The Central Processing Unit is a processing device that performs typical control
function. Bitcoin mining was first performed using general-purpose CPUs of simple
computers.
2. GPU: The Graphic Processing Unit is the processor that handles display functions and is
generally referred to as graphic card. GPU-based Bitcoin mining quickly took over CPU
mining in late 2010 because it offered superior efficiency and processing speed.
3. FPGA: A Field Programmable Gate Array is a particular hardware device whose
performance is significantly superior to graphic cards and comes close to performance
of customised hardware chips. FPGA-based Bitcoin mining gained significant traction
in mid-2011 when the first hashers switched from GPUs to FPGAs.
4. ASIC: An Application-Specific Integrated Circuit is a customised hardware chip
specifically optimised for performing a single task. The emergence of the first Bitcoin
ASICs in mid-2012 rapidly displaced FPGA-based mining
Data from an augmented sample of more than 30 cryptoasset mining hardware manufacturers show
that nearly 68% of manufacturers produce ASICs or at least components that are application-specific
machines optimised for a particular mining algorithm (Figure 43). GPU-based rigs are manufactured by
roughly one third of hardware producers. 71% of surveyed manufacturers indicate that they specialise in
the production of one type of equipment.
Figure 43: ASICs and GPU rigs are the most produced mining hardware equipment
Note: the analysis is based on a sample of 30+ manufacturers obtained from a combination of survey data and publicly available
data.
Supported Hardware Types
Share of Manufacturers
FPGA
GPU
ASIC
CPU
68%
32%
3%
11%
SECTION 7: MINING SEGMENT
74
However, the majority of selected manufacturers produce custom mining rigs that are composed
of existing hardware machines by other manufacturers. This means that they primarily focus on
repurposing existing components to build fully-integrated hardware optimised for cryptoasset mining.
Only a small number of surveyed manufacturers produce their own chips.
The majority of manufacturers produce custom mining rigs
Every mineable cryptoasset uses a particular mining algorithm for its PoW: some are using the same
algorithm whereas others intentionally choose to use a less common algorithm. Each algorithm has its
own peculiarities and as a result, different types of hardware equipment that are best suited for the
task. This means that a machine that is doing well in Bitcoin mining is not necessarily best suited for
Ethereum mining, and vice-versa.
Over time, increasing popularity and subsequent price increases of certain cryptoassets have spurred
investment into R&D to build application-specific hardware that implements various optimisation
techniques designated specifically for the underlying mining algorithms. This has resulted in the
emergence of ASICs that are optimised for a given mining algorithm: these machines are considerably
more efficient and performant for solving that specific mining algorithm compared to general-
purpose hardware. As long as there is a financial incentive (i.e. high profit margins for mining a specific
cryptoasset), manufacturers will continue designing new chips optimised for the underlying mining
algorithm.
Figure 44: Ethereum's Ethash is the most often supported algorithm by manufacturers
Note: the analysis is based on a sample of 30+ manufacturers derived from a combination of survey data and publicly available
data.
Data suggests that Ethash, a mining algorithm first used by the Ethereum network, is the most often
supported algorithm by hardware manufacturers (Figure 44). Ethash is closely followed by SHA-256
(used in Bitcoin and Bitcoin Cash, among others) and Equihash (used in ZCash) in terms of support by
selected manufacturers.
Supported Mining Algorithms
Share of Hardware Equipment
Equihash
SHA-256
Ethash
46%
40%
40%
CryptoNight
26%
X11
14%
Scrypt
17%
Other
20%
2nd Global Cryptoasset Benchmarking Study
75
Two-thirds of surveyed manufacturers indicate that equipment is primarily produced in a single country,
whereas others report to have production facilities in two or three countries. Reported production sites
are primarily located in China and Taiwan, but South American countries (e.g. Chile, Paraguay), Western
European countries (e.g. France, UK) and Eastern European countries (e.g. Russia, Belarus) were also
mentioned.
Distribution Channels
Mining equipment can be primarily purchased via online stores (71% of surveyed manufacturers) that are
either directly operated by the manufacturers themselves or via third-party websites. 57% of surveyed
manufacturers also conduct direct B2B sales to distribute their equipment directly to customers.
Customers can be professional hashing facility operators (e.g. proprietary hashers, cloud mining service
providers, remote hosting service providers) as well as individuals.
Major hardware manufacturers have indicated that their customer base has experienced tremendous
growth since they launched activities: two manufacturers report growth figures as high as 667% and
2,900%, respectively, for the period running from 2015 to 2017. Interestingly, demand does not seem
to have slowed down significantly in the first two quarters of 2018 despite the downturn in cryptoasset
prices.
Online stores appear to be the main distribution channel for manufacturers
The two largest Bitcoin ASIC hardware manufacturers alone have sold more than a combined 5 million
machines worldwide.51 The total number of direct customers is estimated to be a lower six-digit figure.52
It is worth mentioning that there are millions of users who buy mining contracts at cloud mining service
providers in order to "virtually" mine their preferred cryptoassets. However, they cannot be considered
"hashers" themselves since they do not operate the machines.
It appears that the majority of customers from hardware manufacturers are based in Asia-Pacific
(principally in China), although the share of European and North American customers has been
increasing lately. While the two largest Bitcoin ASICs manufacturers indicate that the majority of
customers are domestic, efforts towards geographically diversifying their customer base beyond
domestic markets can be observed.
How Concentrated Is Manufacturing?
There are more than 30 identified cryptoasset hardware manufacturers worldwide, although the total
number is likely higher. They are present around the world with the exception of South America with
the largest presence in Asia-Pacific and North America (Figure 45).
51
This data point is based on information included in the Bitmain and Canaan pre-IPO prospectuses, which are publicly available.
52
Idem
SECTION 7: MINING SEGMENT
76
Figure 45: Mining hardware manufacturers are primarily based in Asia-Pacific and North America
Note: the analysis is based on a sample of 30+ manufacturers derived from a combination of survey data and publicly available
data.
When looking at the type of mining hardware devices, it appears that 62% of ASIC producers are
based in Asia-Pacific, among which 73% are based in China, whereas 14% and 19% are based in Europe
and North America, respectively. In comparison, 50% of GPU producers are based in North America,
primarily in the United-States, while the remaining are split across Asia-Pacific (25%), Europe (13%) and
MEA (13%).
Mining hardware manufacturing is relatively concentrated
However, as explained above cryptoassets use different mining algorithms which require different
machines for the most efficiency. Mining hardware manufacturing concentration varies from one
cryptoasset (or rather mining algorithm) to another and is dependent on a variety of factors. Hardware
manufacturing for certain algorithms (e.g. Bitcoin's SHA-256) is currently very concentrated (both
geographically and from an operator perspective): small miners raise concerns over growing difficulties
in accessing state-of-the-art hardware equipment in a timely fashion, an issue that large miners seem to
be less concerned by.
Balancing out the current concern about concentration are recent statements from established
semiconductor corporations announcing to enter cryptoasset hardware manufacturing.53 This would
then lead to a greater diversity in hardware offering and reduce concentration in the market. On the
other hand, many cryptoassets can be mined using general-purpose hardware that is easily available and
can be repurposed. Some believe that mining hardware will soon become commoditised: the anticipated
53
For an example, see Samsung (2018) Samsung Electronics Announces Fourth Quarter and FY 2017 Results. Press Release.
Available at: https://news.samsung.com/global/samsung-electronics-announces-fourth-quarter-and-fy-2017-results [Accessed:
02 December 2018].
Geographic Distribution of Mining Hardware Manufacturers
Asia-Pacific
47%
Middle East
and Africa
6%
North America
31%
Europe
16%
2nd Global Cryptoasset Benchmarking Study
77
end of significant breakthroughs in chip design combined with the entrance of new players will likely
reshuffle market dynamics sufficiently and reduce hardware manufacturing concentration overall.
7.3 Mining Facilities
Meet the Hashers
The process of hashing is permissionless: anyone in the world can at least in theory enter cryptoasset
mining by downloading and running mining software on a machine. Since the puzzle difficulty of many
valuable PoW blockchains has risen substantially over time, hashing now often requires the use of many
specialised machines to make operations economically viable. Hashers range from individual hobbyists
running a few rigs at home to large companies operating large-scale data centres hosting 100,000 rigs or
more.
Hashers can mine for their own account (proprietary) or use their data centres to provide services to
customers such as remote hosting (operating and maintaining customer-owned hardware) and cloud
mining (renting hashpower). They generally point their hashpower to a mining pool to smoothen pay-outs.
Facility Set-up Decision Factors
Surveyed miners were asked to rank the most important decision factors used for assessing the
suitability of a location for a new mining facility (Table 8 in Appendix). The following five factors stand
out:
1. Access to ample and low-cost electricity supply
Running hashing facilities requires substantial amounts of electricity that can be adjusted
dynamically in accordance with market conditions. It is not unusual for mining farm operators
to make deals with local energy suppliers to guarantee access to sufficient and affordable
electricity.54
2. Friendly regulatory environment
Local geographies that take a favourable regulatory stance with regards to mining and create
incentives (e.g. tax-related, subsidies) will attract hashers.
3. Stable political situation
Ranked third, it highlights the need for a stable and predictable political environment that
protects property rights and has a functional justice system. Small miners indicate that the
possibility of government seizure or shutdown of their facilities constitutes the biggest risk.
4. Good Internet connectivity
A fast and reliable Internet connection is paramount for hashers in order to quickly receive and
broadcast data (e.g. instructions, pool shares proving the "work" performed).
5. Cold climate
Mining machines consume a lot of energy, which requires constant cooling to prevent them for
overheating. Facilities located in regions with cold climate offer substantial advantages in terms of
cooling cost savings.
Interestingly, the presence of cheap land and skilled labour was ranked much lower, suggesting that these
are only secondary factors with minor impact on decision-making. Only South American miners indicated
that a low crime rate would be a major decision factor.
54
In fact, large surveyed miners ranked sudden increases in electricity prices as the highest operational risk factor (see Table 8 in
Appendix).
SECTION 7: MINING SEGMENT
78
Distribution of Mining Facilities
Through a combination of survey data and public sources, the research team built a repository of 128
mining facilities around the globe.55 Figure 46 illustrates that hashing activities have scaled globally and
are distributed across all world regions, albeit not equally: some continents (e.g. Asia-Pacific, North
America) host significantly more mining activity than others (e.g. MEA).
Figure 46: The Global Cryptoasset Mining Map
55
Data was collected from company websites, press releases, news articles, public forums, social media, and insights from industry
experts. When available, the research team aggregated the capacity of regional facilities measured in megawatt (MW) and
estimated the level of mining activity in each country. Countries hosting more than 40 MW identified in the above analysis are
considered to have a high level of activity. Countries with low activity levels are known to host mining facilities, whereas countries
with suspected activities could not be confirmed by multiple sources. In total, 1,745 MW powering mining facilities have been
identified.
Geographic Distribution of Mining Facilities
British Columbia
Washington State
Alberta
Quebec
New York State
Canada
Brazil
United States
200MW
100MW
50MW
Identified megawatts (MW)
Level of activity
High
Low
Suspected
No indication
2nd Global Cryptoasset Benchmarking Study
79
Sweden
Hebei
Xinjiang
Inner Mongolia
Sichuan
Armenia
Georgia
India
China
Russia
Australia
Note: this map is based on a dataset of 128 mining facilities. A total capacity of 1.7 gigawatts (GW) could be identified for 93
facilities. High-activity countries have 40 megawatts (MW) or more identified capacity; low-activity and suspected activity levels
are estimates.
SECTION 7: MINING SEGMENT
80
The mining map shows that hashing facilities are primarily located in China, North America (USA
and Canada), and North-Eastern Europe (Russia and Georgia). Facilities located in China have been
responsible for the majority of mining in the past and the country remains a major hub despite changes in
the regulatory environment.
In fact, the government's restrictive measures on mining seem not to have led to a massive exodus of
Chinese miners overseas. While some large Chinese miners have opened additional facilities overseas,
small domestic miners have been disproportionately affected by falling prices and some were forced to
shut down their facilities. The bulk of mining is located in the Northern provinces (e.g. Inner Mongolia,
Xinjiang, Hebei, Heilongjiang) and Southwest China (e.g. Sichuan, Yunnan, Guizhou).
Substantial growth of mining activities can be observed
in North America, but China remains a major hub
Specific states in the USA (e.g. Washington, New York), Canadian provinces (e.g. Qubec, British
Columbia, Alberta) and some Scandinavian countries (e.g. Iceland, Norway, Sweden) seem to particularly
benefit from these developments, as rapid growth in local mining activities can be observed.56
It is no coincidence that these regions satisfy all five criteria laid out in the previous subsection: cheap
electricity is provided in ample volumes often from excess capacity of hydroelectric or geothermal
power, the seasonal cold climate drives down cooling costs, and the countries are among the most
developed in the world (i.e. having very high human development index scores, good network
infrastructure, and functioning institutions).
A growing level of activity can also be observed in some South American countries (e.g. Argentina,
Colombia, Venezuela) and Western European countries (e.g. France, UK, Switzerland). It is worth
noting that the mining map is incomplete and does only capture a share of global mining activities: some
countries may have higher albeit hidden activity levels.
Nevertheless, available data suggests that cryptoasset hashing has become more geographically
distributed since 2017, particularly with China losing relative "market share" to some North American
and Scandinavian regions. Improvements in the geographic distribution will make it more difficult
for attackers to seize or shut down hashing facilities, a concern that was ranked first by small miners
that took part in the survey. Similarly, it will make global mining less dependent on local events, both
economically and politically-driven (e.g. sudden increases in electricity prices, which is the highest-ranked
concern of large miners, or the establishment of a tax regime on mining profits particularly worrying
small miners and individuals).
A relatively large number of operators cannot hide the fact that a few are dominating
Similarly, hashrate ownership appears to be relatively distributed across many operators: the research
team could identify more than 60 different facility and data centre operators, with the majority of
identified hashers operating a single facility. However, facility size and capacity do vary significantly from
one farm to another. Moreover, hashers can also operate multiple facilities, with some large companies
running more than a dozen.
Overall, concern about geographical concentration of hashpower has been overplayed: in reality, hashing
appears to be globally distributed. However, when it comes to concentration of hashpower in terms of
ownership, the picture is less clear: while there are at least several dozen operators of facilities around
the globe, a small number of large hashers seem to occupy a dominant position.
56
However, this tendency for clustering in specific locations might be not always be harmonious: In Qubec, for instance, the recent
surge in mining activities has led to stricter governmental intervention and higher electricity charges for companies in the mining
industry in June 2018 (Data available at http://publicsde.regie-energie.qc.ca/projets/457/DocPrj/R-4045-2018-A-0001-Dec-
Dec-2018_06_18.pdf [Accessed: 02 December 2018]).
2nd Global Cryptoasset Benchmarking Study
81
How Much Energy Does Cryptoasset Mining Consume?
The energy-intensive nature of PoW mining and its potential negative implications on the environment
have sparked heated debates. Multiple studies estimating the total energy consumption were published
using various methodologies which resulted in widely divergent figures.57
Estimating the actual energy usage of the major PoW cryptoasset systems requires extensive data on the
type of mining equipment used, the nature of energy sources, and the overall energy efficiency of mining
data centres. For the lack of reliable data, the research team bases the estimate on a range rather than a
precise value.
The top-6 cryptoassets consume between 52 and 111 TWh of energy a year
The lower-bound estimate follows the same methodology as last year's study: it assumes that all hashers
are using the most efficient hardware available in the market and run the most efficient data centres.58
The upper-bound estimate, based on a methodology developed by Marc Bevand59, assumes that all
hashers always run the least efficient hardware available in the market as long as the hardware yields a
positive return when purely considering electricity costs (i.e. operational expenses).60
A separate analysis is conducted for the six major PoW cryptoassets (Bitcoin, Bitcoin Cash, Ethereum,
Litecoin, Monero, and ZCash) and outputs are combined to provide an aggregate estimate. Figure 47
shows that as of mid-November 2018, the top-6 cryptoasset networks are estimated to collectively
consume between 52 and 111 terawatt-hours (TWh) of energy a year, up from between 18 and 42 TWh
just a year before. Bitcoin alone accounts on average for 75% of the total energy consumption.
57
For a comparison of different estimates and their underlying methodology, see Kolbie, N. (2017) How much energy does Bitcoin
mining really use? It's complicated. Wired. Available at: https://www.wired.co.uk/article/how-much-energy-does-bitcoin-mining-
really-use [Accessed: 02 December 2018].
58
Using Bitcoin as an example, the following hardware types were used for each respective period as the most efficient equipment:
Antminer S7 (Q1 to Q3-2016), Avalon A721 (Q4-2016), Antminer T9 (Q1 and Q2-2017), Antminer S9 (Q3-2017 to Q2-2018),
DragonMint 16T (Q3 and Q4-2018). Furthermore, it was assumed that facilities have a power usage effectiveness (PUE) similar
to the most efficient data centres in the world (1.03) as well as a low parasitic power consumption of 5%.
59
Bevand, M. (March 2017) Electricity consumption of Bitcoin: a market-based and technical analysis. Personal Blog. Available at:
http://blog.zorinaq.com/bitcoin-electricity-consumption/ [Accessed: 06 November 2018].
60
Using Bitcoin again as an example, the following hardware types were used for each respective period as the least efficient
equipment: Avalon A6 (Q1 to Q4-2016), Avalon A721 (Q1 to Q4-2017), Antminer T9 (Q1 and Q2-2018), Avalon A821 (Q3 and
Q4-2018). Furthermore, it was assumed that facilities have a PUE of 1.33 and a parasitic power consumption of 15%.
SECTION 7: MINING SEGMENT
82
Figure 47: The combined energy consumption of the top-6 PoW coins has consistently grown
despite the recent downturn in market capitalisation
Note: 2018 Q4 estimates are based on available hashrate data from October to mid-November 2018. The recent decline in
hashpower (and as a result total energy consumption) in late November 2018 is not covered by the chart.
Taking the latest mid-point of the estimated range as a reference (82 TWh), it can be established that
the top-6 cryptoasset systems consume approximately as much energy as the entire country of Belgium
in 2016.61 At the same time, this figure amounts to less than 0.01% of the world's total annual energy
production62 , or the equivalent of all electricity generated by biomass and solar energy in Germany
alone.63
Energy consumption is a direct function of hashpower. Unless new, more energy-efficient mining
hardware is introduced, total consumption will rise in a linear fashion with hashpower.64 Skyrocketing
market prices in the second quarter of 2017 led to an exponential increase in hashpower, which in return
triggered a substantial rise in the amount of energy consumed by mining facilities. Data suggest that
hashpower growth is lagging behind market price growth: hashers often cannot immediately increase
production when running at full capacity.
Total energy consumption increased more than fivefold
between mid-2017 and mid-November 2018
Interestingly, total hashpower and as a result energy consumption has continued its steep growth
despite the crash of the cryptoasset market that prompted prices to plummet. However, Bitcoin's
61
CIA World Factbook. Data available at: https://www.cia.gov/library/publications/resources/the-world-factbook/fields/253rank.
html [Accessed: 25 November 2018].
62
IEA Energy Atlas. Data available at: http://energyatlas.iea.org/#!/tellmap/-1118783123 [Accessed: 25 November 2018].
63
Burger, B. (2018) Power generation in Germany: Assessment of 2017. Fraunhofer Institute for Solar Energy Systems ISE. Available at:
https://www.ise.fraunhofer.de/content/dam/ise/en/documents/publications/studies/Stromerzeugung_2017_e.pdf, p.8 [Accessed:
25 November 2018].
64
New-generation hardware with vastly improved energy efficiency is introduced periodically into the market, which can
temporarily lead to a decrease in the total amounts of electricity consumed. However, the increase in margins will incentivise
other hashers to expand operations, resulting in a higher hashrate and ultimately PoW difficulty. Over time, the energy efficiency
effect is cancelled out by the increase in hashrate, and eventually total energy consumption levels rise again.
Estimated Energy Consumption Range
Top-6 PoW Cryptoassets
60
40
20
100
$100
$0
120
$50
$200
$250
$300
$350
80
0
2018 Q4
$150
Total TWh consumedCombined market capitalisation ($billion)2016 Q1 2016 Q2 2016 Q3 2016 Q4 2017 Q1 2017 Q2 2017 Q3 2017 Q4 2018 Q1 2018 Q2 2018 Q3
Aggregate upper bound (lhs)
Aggregate lower bound (lhs)
Bitcoin lower bound
Bitcoin upper bound
Combined market cap (rhs)
Bitcoin range
2nd Global Cryptoasset Benchmarking Study
83
hashrate has since come down from 58 Exahashes per second (Ehs) in early November 2018 to 34 Ehs
at the time of writing in late November 2018. This decline has resulted in a reduced energy consumption,
demonstrating that cryptoasset mining is a dynamic process that is constantly self-adjusting.
Misperceptions About Energy Cost Per Transaction
Many energy consumption estimates include comparisons with traditional payment
systems based on transaction throughput. The resulting energy cost per transaction is not a
meaningful metric in the context of PoW blockchains for the following reasons:
Throughput unrelated to energy consumed: the level of energy required for the
networks to function is independent from the number of processed transactions.
Hidden semantics: a single blockchain transaction can include thousands of payments,
settle layer-2 network transactions (e.g. open and close channels in the Lightning
network), and represent potentially billions of timestamped data points.
Different value proposition: unlike traditional payment systems, PoW blockchains
are designed to function as censorship-resistant value transfer systems. This value
proposition requires engaging different trade-offs that result in substantial operational
costs.
How Wasteful Is Cryptoasset Mining?
Some studies have attempted to estimate the environmental impact of cryptoasset mining on our
planet. Krause and Tolaymat (2018) estimate that Bitcoin, Ethereum, Litecoin and Monero have been
responsible for 3-15 million tonnes of CO2 emissions since January 2016,65 whereas Mora et al. (2018)
project that Bitcoin alone could produce sufficient CO2 emissions to push global warming above 2C
within less than three decades.66
These analyses neglect the mining energy mix the nature of the energy sources used to operate the
hashing facilities. For instance, the energy footprint of one MW of energy generated by a coal-fired
power station is not equivalent to the footprint of one MW of energy generated by a hydroelectric
power station particularly if the station produces overcapacities that are unmet by traditional demand.
The environmental impact of cryptoasset mining
is dependent on the energy mix used
Using the mining facility dataset previously introduced, it is possible to conduct a simple analysis of the
average energy mix used by identified hashers. Contrary to popular beliefs, cryptoasset mining does
not exclusively rely on fossil fuels: in fact, more than half of hashing facilities run on an energy mix that
contains a share of renewables (Figure 48).
65
Krause, M. J., and Tolaymat, T (2018) "Quantification of energy and carbon costs for mining cryptocurrencies" Nature Sustainability.
Available at: https://www.nature.com/articles/s41893-018-0152-7.pdf. [Accessed: 30 November 2018].
66
Mora, C., Rollins, R.L., Taladay, K., Kantar, M.B., Chock, M.K., Shimada, M., and Franklin, E.C. (2018) "Bitcoin emissions alone could
push global warming above 2 degree" Nature Climate Change. Available at: https://www.nature.com/articles/s41558-018-0321-8.
pdf. [Accessed: 30 November 2018].
SECTION 7: MINING SEGMENT
84
Figure 48: Identified mining facilities often run on an energy mix that contains a share of
renewables
Note: data is based on a dataset of 128 hashing facilities around the globe. Megawatt figures are available for 93 facilities.
Less than a quarter of identified miners do not use any forms of renewable energy sources at all,
although the energy mix of one quarter of facilities could not be identified.67 Certain regions such as
Xinjiang Province in China rely almost entirely on coal. Nevertheless, weighing the farms by identified
megawatts results in a similar picture and shows that cryptoasset mining is much less dependent on fossil
fuels than anticipated.
Identified facilities draw on average 28% of their energy requirements from renewables
However, the share of renewables varies considerably from one facility to another: while some only use
a marginal proportion, others run almost exclusively on renewables. On average, roughly 28% of the
total energy supply for both small and large facilities is generated through renewable sources. Among
renewables, hydroelectric power is the most frequently used energy source. Nearly half of the identified
megawatt capacity featured in the cryptoasset mining map is generated through hydropower. It is worth
noting that the mining map identifies 30% of the lower-bound total energy consumption estimate.68
An interesting pattern emerges when comparing the energy mix with the location of the respective
mining facilities. Regions with substantial green power sources seem to become attractive targets
for miners, since these locations tend to overlap with places where there is an abundance of low-cost
hydroelectric power that is unused and stranded. Energy in these locations is often cheap because
demand cannot compensate the oversupply.
67
It should be noted that mining facilities running on renewables may be more incentivised to publicise and advertise their energy
mix, whereas facilities running exclusively on fossil fuels have less incentive in sharing information.
68
The location of many mining facilities remains unknown; the 1.7 GW figure should thus be considered as a lower-bound estimate.
For instance, mining activity in China is expected to be much higher than reported.
Share of Renewable Energy in the Energy Mix
by Identified MW
Total
1,745 MW
26%
18%
56%
Energy mix includes renewables
Energy mix does not include renewables
N/A
Share of Renewable Energy in the Energy Mix
by Mining Farm
Total
128 mining farms
31%
9%
60%
2nd Global Cryptoasset Benchmarking Study
85
Hydroelectric power is the most used renewable energy source among identified facilities
The mining map confirms the pattern and shows that mining activities tend to increasingly cluster and
congregate in locations with excess capacities in renewables (in particular hydroelectric power), such as
for instance Western and Southwestern China, the North-East and North-West of the USA, as well as
South-Eastern Canada and Iceland. While the claims of some proponents suggesting that cryptoasset
mining will drive the "green revolution" may be a bit far-fetched, it is fair to say that regions with
abundant renewable energy sources progressively attract miners chasing low-cost electricity.
Some renewable energy sources require the use of fossil fuels in their production, whereas others can be
very expensive to produce. Moreover, renewables such as hydro (seasonal variances, dry periods), wind
(weather-dependent), and solar (available only for a limited number of hours per day) are intermittent:
supply is subject to seasonal changes and conditions, and often need to be supplemented by alternative,
non-renewable energy sources during certain periods.
However, as such renewable energy sources fluctuate in their production, they can also overproduce
relative to local demand. Cryptoasset mining may soak up local overcapacities and prevent the waste of
otherwise unused renewable energy power that cannot be easily stored and transmitted over large
distances. However, if such overcapacities are less than the demand of mining operations, this may create
an increase in local energy prices and hurt local businesses.
It is important to consider the energy mix used in the hashing process as well as potential alternative uses
(or the lack thereof) when assessing the impact of cryptoasset mining on global carbon emissions. The
preceding analysis concludes that concerns over Bitcoin and PoW cryptoassets in general directly
and significantly contributing to climate change are largely overestimated at the time of writing.
What Do Miners Think?
Miners appear to be relatively indifferent with regards to their energy mix and whether it contains
renewables. Instead, they prioritise low-cost electricity and a steady, reliable energy supply. As
previously seen, it turns out that in some cases, their quest for cheap and ample electricity drives them
to locations where excess capacities from renewables have significantly driven electricity prices down.
When asked about their views on the environmental impact of PoW mining, surveyed miners provided
similar responses to the 2017 survey: while most acknowledge that mining has an environmental impact
on the planet to some degree, they urge commentators and critics to put energy consumption into
perspective.
While acknowledging the environmental impact of PoW mining,
most miners think it needs to be put into perspective
Many comment that other industries, such as physical commodities mining, use substantial amounts of
energy as well and are not necessarily known for having a green footprint. Some miners also question
the utility of the existing financial system and thus the electricity required to make this system function.
Furthermore, both small and large miners strongly believe that the negative environmental externality
might be alleviated by switching to more environmentally-friendly power sources. The previous analysis
seems to confirm this trend to some extent.
A change in the PoW algorithm or a move from resource-intensive PoW mining to less energy-intensive
PoS mining to alleviate the environmental impact do not seem to be acceptable options to most miners.
This is not surprising as miners have made considerable investments into specialised equipment
and facilities that they need to recover. It will be interesting to see how the planned move by some
cryptoassets from PoW to PoS will affect miners, and whether they will grow into a new role as "stakers".
SECTION 7: MINING SEGMENT
86
7.4 Pool Operators
Pool Operations
A mining pool is a structure that "pools" together computational resources provided by connected
hashers (pool contributors) in order to increase the likelihood and frequency of finding a new block,
which results in smoother pay-outs. A pool operator creates a candidate block and sends the template
to contributors, who will then perform hashing until a new block is found. Contributors send back pool
shares to the pool operator who will distribute mining rewards in proportion to the amount of work
performed.
The requirements for running pools are generally low: operators preferably need to run a fully-validating
node of the cryptoasset system they are mining, maintain a simple server, and have a fast and stable
Internet connection. Pools can range from single-operator projects run by hobbyists to sophisticated
service providers with dedicated technical and customer support.
Pools generally do not restrict membership to users from specific jurisdictions
Pools appear to have no restrictions with regards to accepting contributors: only one out of eight
surveyed pools mentioned that they are a private pool for investors and do not accept investors from
specific regions. However, the majority of pools do implement policies to detect botnet activity and
prevent distributed denial-of-service (DDoS) attacks.
On average, more than half of registered pool
members actively contribute at least once a week
The number of contributors varies widely from one pool to another, as does the share of active members.
Most pools consider hashers to be active when they contribute hashpower at least once a week.
According to survey data, the share of active members averages 51%, but can range from a mere 2% up
to 100%.
Data suggests that hashpower contribution follows a power law distribution: on average, one third of
the pool's total hashrate is provided by the top-1% of contributors, whereas 10% of active pool members
contribute 68% (Figure 49). Nevertheless, figures are widely divergent: some pools indicate that up to
70% of their total hashrate is supplied by the top-1% of their members, whereas others are much more
distributed in that only 30% of the total hashrate is provided by 10% of the pool members.
Figure 49: A small share of pool members contribute the majority of total pool hashrate
Share of Total Pool Hashrate
Average
Median
70%
68%
Top 10%
Top 1%
20%
33%
Top 5%
55%
58%
2nd Global Cryptoasset Benchmarking Study
87
Pool Concentration
A repository of 147 identified mining pools across the cryptoasset ecosystem was built using a
combination of survey data and public sources. Data shows that the vast majority (92%) of identified
operators manage only a single pool, which seems to challenge the common view that mining pools are
too concentrated (Figure 50). Interestingly, however, pool operators who responded to the survey
(mostly large miners) are much more likely to operate multiple pools: in fact, more than one third
indicates to run five or more pools.
Figure 50: While operators generally tend to manage a single pool, three-quarters of survey
participants run two or more pools
Note: the survey sample is based on data from survey participants. The augmented sample is based on a list of 147 identified
mining pools across the cryptoasset ecosystem.
Data suggests that there is no clear pattern with regards to the pool decision-making process, but that
pool policy is relatively diverse. Changes to pool policy (e.g. decision to mine a new coin) are equally likely
to be made unilaterally by a single individual (38% of surveyed pools) or a group of operators (38%). The
remaining 26% of pools take a more user-focused approach: users can participate in a vote-by-CPU (or
equivalent) agreement. Other mining pools indicate using a combination of all these factors to instigate
modifications of the pool policy.
More than a third of surveyed pools are fully controlled by a single person
The geographic distribution of mining pools varies significantly from one cryptoasset to another.
However, they share in common that Asia-Pacific, Europe, and North America are generally dominating,
albeit in different constellations (Figure 51). As of mid-November 2018, Bitcoin pools seem to be
relatively equally distributed between the three aforementioned regions, whereas the Bitcoin Cash pool
landscape seems to be dominated by pools located in Asia-Pacific. European pools appear to be dominant
in Ethereum and Monero mining.
Number of Mining Pools by Operator
Share of Pool Operators
Survey sample
Augmented sample
3 Pools
4 Pools
5+ Pools
37%
25%
13%
2 Pools
1%
2%
4%
1%
Single Pool
25%
92%
SECTION 7: MINING SEGMENT
88
Figure 51: The mining pool landscape can vary a lot from one cryptoasset to another
Note: in some cases, insufficient information requires the geographic location of operators to be determined by server location.
Data as of mid-November 2018.
It is worth noting that determining pool location can be tricky: some operators remain anonymous,
whereas other have servers distributed across the world. Renting servers in different countries and
regions is relatively simple and means that pool operators can quickly move operations between
different locations.
Ethereum has the highest number of identified pools (87), followed by Bitcoin (57) and Zcash (45).
However, the total number of pools is not a relevant metric in itself: instead, the relative share of
hashpower controlled by each pool needs to be taken into account.69 As of mid-November 2018, only
four pools combined control more than 50% of Bitcoin's total hashrate. The picture looks even more
concentrated for other cryptoassets: it would only take three pools to collude in order to perform a
51%-attack on Bitcoin Cash, Ethereum, Litecoin and the Monero network; whereas two ZCash pools
alone control more than half of the network's hashrate.
A minority of pools control the majority of hashrate for the major cryptoassets
69
Colluding miners that collectively control more than 50% of a cryptoasset network's hashrate will on average find more blocks
than their honest competitors. This enables them to rewrite transaction history and perform double-spends.
Asia-Pacific
21%
Europe
40%
Unknown 2%
North
America
37%
ZCash
45 identified pools
2 pools control >50% hashrate
Asia-Pacific
16%
Europe
47%
Unknown 3%
North
America
34%
Monero
38 identified pools
3 pools control >50% hashrate
Asia-Pacific
52%
Europe
20%
North
America
28%
Litecoin
26 identified pools
3 pools control >50% hashrate
Asia-Pacific
58%
Europe
9%
North
America
33%
Bitcoin Cash
13 identified pools
3 pools control >50% hashrate
Unknown 1%
Asia-Pacific
29%
Europe
49%
North
America
21%
Ethereum
56 identified pools
4 pools control >50% hashrate
Asia-Pacific
44%
Europe
25%
North
America
29%
South America 2%
Bitcoin
56 identified pools
4 pools control >50% hashrate
2nd Global Cryptoasset Benchmarking Study
89
While at first sight it appears that the mining pool landscape is significantly concentrated, a second
look reveals that pool operators have less influence than anticipated. Hashers generally want to ensure
that the pool they contribute to is engaging in a behaviour that is in agreement with their philosophy
towards the chosen cryptoasset. In the event of disagreement or unstable reward allocation, miners are
free to switch pools. Past events have shown that this also applies when a particular pool becomes too
dominant: in 2013, Bitcoin mining pool GHash.io reached more than 50% of the total hashrate for a short
period, which prompted hashers to proactively point their hashpower to another pool in order to avoid
potential harmful concentration.
Low switching costs act as a check and
balance system on pool operator behaviour
Assuming low switching costs, the ability of hashers to quickly move from one pool to another represents
an implicit threat that acts as a check and balance system on mining pool behaviour. This dynamic and
self-regulating process has worked relatively well so far, with no major blockchain reorganisation attacks
observed for the dominant PoW cryptoassets.
FUTURE OUTLOOK
90
FUTURE OUTLOOK
Sentiment questions from survey data suggest that the trend towards multi-coin support is likely to
continue: all small storage service providers envisage to support more cryptoassets when asked about
their cryptoasset roadmap. Only less than a quarter of large storage service providers indicated not
planning to support more cryptoassets (Figure 52).
Figure 52: All single-coin wallets plan to support more cryptoassets; small wallets more likely to
add more cryptoassets than large wallets
Regardless of the scale of their activities, the majority of storage service providers who plan to support
more cryptoasset have three or more cryptoassets on their roadmap. This trend has accelerated: while
only half of storage service providers surveyed in 2017 mentioned plans to support more cryptoassets,
91% of storage providers have more cryptoassets on their roadmap in 2018. Similarly, it is likely that the
trend towards multi-segment services is going to continue.
Non-cryptocurrency cryptoassets ("cryptotokens") became more popular in the ecosystem, primarily
driven by the wide adoption of the ERC-20 standard on the Ethereum network. This led to a boom in
token-based fundraising and a flurry of Initial Coin Offering (ICO) activities globally. The ICO market will
be examined in detail in a future report. The increase in interest and subsequent usage of cryptoassets
brought into the foreground limitations of base layer scaling and led to the launch of so-called "layer-2
solutions", such as the eagerly-awaited Lightning Network on Bitcoin.70
Service providers were also asked about their views on new developments in the cryptoasset ecosystem
and how these are expected to impact their business models and operations (Table 5). Within a
remarkably large ranking range, off-chain "layer-2" payment solutions are perceived to have the greatest
impact within the coming 12 months. Payment-only companies in particular support this view.
70
During the height of the boom, the Bitcoin blockchain experienced significant delays in processing transactions, with average fees
rising to levels above $50. Similarly, the Ethereum blockchain was clogged for a few days because of one single gaming application
that suddenly became popular (CryptoKitties). Layer-2 solutions refer to a variety of techniques that aim to materially increase
transaction speed and throughput as well as substantially decrease transaction costs by moving payments off-chain.
Cryptoasset Roadmap
Share of Storage Service Providers
6+ cryptoassets
1 cryptoasset
3-5 cryptoassets
No cryptoassets
2 cryptoassets
21%
5% 5%
11%
Large storage providers
16%
6%
29%
48%
58%
Small storage providers
Multi-coin support
Single-coin support
53%
7%
22%
9%
9%
40%
40%
20%
Number of cryptoassets to be added
2nd Global Cryptoasset Benchmarking Study
91
Table 5: Innovations in off-chain payment networks ("layer-2") thought to have the largest impact
on service providers' business model and operations
Impact
Payment-only
Exchange-only
Storage-only
Multi-segment
Stablecoins
3.22
2.90
3.73
3.34
CBDC
3.00
2.76
3.18
2.72
Non-fungible tokens
2.40
2.84
2.82
2.69
Security tokens
2.50
3.33
3.36
3.40
Layer 2
4.30
3.65
4.00
3.98
On-chain scaling
3.60
3.11
3.09
3.70
Other
1.00
1.00
3.00
3.40
In contrast, on-chain scaling is perceived to have a lower impact, principally for payment-only and multi-
segment companies. It is expected that in vogue technologies like stablecoins will significantly impact
operations of storage-only and multi-segment firms. Central bank-issued digital currency (CBDC),
however, is not seen as having a large impact in the coming year.
Independent of industry segment, non-fungible tokens (tokens that are cryptographically-unique, for
instance implemented by standards such as Ethereum's ERC-721) currently rank as the least impactful
innovation for future evolution. Finally, exchange-only, storage-only, and multi-segment firms expect
security tokens tokens representing a traditional security agreement to play a more significant role in
the future.
Yet, all these future plans were made during or shortly after the 2017 cryptoasset boom and may
represent a future that is now beyond the reach of the cryptoasset industry. The following analysis
shows that this pattern of rapid-expansion and collapse has characterised the cryptoasset ecosystem
since Bitcoin emerged in 2009.
We define a bubble as the market capitalisation of a cryptoasset appreciating by a multiple of 10 or more
within a period of 6 months or less, followed by a substantial decline.71 Figure 53 shows the evolution of
the market capitalisation of five leading cryptoassets (bitcoin, ether, ripple, bitcoin cash, and litecoin) and
illustrate how a series of speculative bubbles can be observed for each of the selected cryptoassets.
71
Using price instead of market capitalisation will result in different dates on the bubbles, even when using the same definition. The
reason for this discrepancy is rooted in the different supply schedules implemented by each cryptoasset.
Respondents scored these categories on a 1-5 scale:
1: Not important at all 2: Not important 3: Neutral 4: Somewhat important
5: Very important
Lowest average score
Highest average score
FUTURE OUTLOOK
92
Figure 53: Cryptoasset markets have seen a succession of local and global ecosystem bubbles since
their inception
Note: data sourced from CoinMarketCap and Coin Dance.
Cryptoasset market bubbles can be categorised into local and global ecosystem bubbles. Local ecosystem
bubbles only affect a particular cryptoasset and its ecosystem but are largely isolated from the overall
development of the global cryptoasset market. In contrast, global ecosystem bubbles refer to market
frenzies that affect the majority of cryptoassets and the entire ecosystem as a whole.
Before 2013, the cryptoasset ecosystem was dominated by Bitcoin and therefore market bubbles were
mostly limited to local Bitcoin bubbles. An example is the "Great Bubble of 2011" that saw the BTC price
peak at nearly $32 in June 2011 before declining again in value by more than 90% over a period of four
months.
Local ecosystem bubbles occur on a regular basis: XRP's price rapidly increased and collapsed in both
2014 and 2015, whereas ETH saw its market capitalisation rise sharply in early 2016 after a successful
launch several months before.
Local ecosystem bubbles are much more common than global ecosystem bubbles
The first global ecosystem bubble built up in late 2013 and led to a more than ten-fold increase in
the aggregate cryptoasset market capitalisation following the announcement that China would allow
cryptoasset trading and the time of Mt. Gox trading bots. It began deflating in early 2014 when China
rolled back their announcement and Mt. Gox was hacked.
Evolution of the Cryptoasset Market Capitalisation
$10
$1
$100
$1,000
$10,000
$100,000
$1,000,000
Total market capitalisation in $million (log scale)Nov-10
No
v-
14
Nov-12
No
v-
16
Nov-11
No
v-
15
No
v-
13
No
v-
17
Fe
b-11
Fe
b-
15
Fe
b-13
Fe
b-
17
Fe
b-12
Fe
b-
16
Fe
b-
14
Fe
b-
18
Aug-11
Au
g-
15
Aug-
13
Au
g-
17
Aug-12
Au
g-
16
Au
g-
14
Au
g-18
M
ay
-1
1
M
ay
-1
5
M
ay
-1
3
M
ay
-1
7
M
ay
-1
2
M
ay
-1
6
M
ay
-1
4
M
ay
-1
8
Nov-18
2011
2012
2013
2014
2015
2015
2016
2013-14
2017-18
Bitcoin (BTC)
Ethereum (ETH)
Ripple (XRP)
Litecoin (LTC)
Bitcoin Cash (BCH)
Local ecosystem bubble
Global ecosystem bubble
2nd Global Cryptoasset Benchmarking Study
93
All previous global bubbles were dwarfed by the market frenzy that began in April 2017. Fueled by
intense media coverage, the tempting promise of 'get-rich-quick' ICO schemes, and an oversupply of
new coins and tokens, many first-time retail and institutional cryptoasset investors rushed in. Aggregate
cryptoasset market capitalisation exploded more than 25-fold to peak at nearly $800 billion, before
rapidly declining throughout the year 2018.
The collapse in prices, and subsequent media coverage of the losses borne by speculative investors in
2018, created a media narrative in which Bitcoin, cryptoassets, and ICOs were not only declared bubbles
but also declared dead.
Statements proclaiming the death of the cryptoasset industry have been made after every global
ecosystem bubble. While it is true that the 2017 bubble was the largest in Bitcoin's history, the market
capitalisation of both Bitcoin and the cryptoasset ecosystem still exceeds its January 2017 levels -
prior to the start of the bubble. This report has shown that the speculation of the death of the market
and ecosystem has been greatly exaggerated, and so it seems likely that the future expansion plans of
industry participants will, at most, be delayed.
APPENDIX: SENTIMENT QUESTIONS
94
APPENDIX: SENTIMENT QUESTIONS
The survey asked participants several sentiment questions. In general, industry participants are more
concerned about operating risk factors in 2018. It is further noted that small and large firms have largely
begun to converge in their perceptions and sentiments.
Table 6: Exchanges rank major operational risks
IT
Security
Fraud
AML/KYC
Enforcement
Regulatory
Burden
Risks
Competition
Negative
Publicity
Bank
Relationship
Entering Bank
Relationship
Lack of
Talent
Large
2018
4.20
3.83
3.40
3.84
3.29
3.52
3.54
3.63
3.83
2017
3.17
2.08
2.75
3.50
2.58
2.75
2.67
2.67
2.33
Small
2018
3.81
3.48
3.17
3.78
3.21
3.30
3.48
3.69
3.48
2017
3.93
3.50
2.64
2.89
3.00
2.93
3.79
3.79
2.52
Table 7: Miners rank major concerns over operational risks
Operational Risks
Small Miners
(incl. Individuals)
Large Miners
2017
2018
2017
2018
Sudden increase in energy prices
N/A
3.09
N/A
3.54
Intensive competition among miners of the same cryptoasset
3.17
3.33
3.30
3.23
Cyber attacks (e.g. DDoS)
2.77
3.07
3.00
3.31
Lack of immediate availability of state-of-the-art hardware
2.94
3.35
2.40
2.46
Declining popularity of the cryptoasset you mine
N/A
2.95
N/A
3.00
Unexpected change to protocol
2.52
3.00
1.64
3.38
Increased taxation of mining profits
N/A
3.16
N/A
2.85
Regulations creating barriers to mining
N/A
3.31
N/A
2.92
Government seizure or shutdown of your mining-supporting facilities
N/A
3.55
N/A
2.38
Respondents scored these categories on a 1-5 scale:
1: Completely disagree 2: Disagree 3: Neutral 4: Somewhat agree 5: Completely agree
Lowest average score
Highest average score
Respondents scored these categories on a 1-5 scale:
1: Not concerned at all 2: Not concerned 3: Neutral 4: Somewhat concerned 5: Very concerned
Lowest average score
Highest average score
2nd Global Cryptoasset Benchmarking Study
95
Table 8: Miners rank major decision factors for choosing a location for establishing new hashing
facilities
Assessment Factors for Setting up a New Mining Facility
Small Miners
(incl. Individuals)
Large Miners
Stable political environment
4.37
4.63
Friendly regulatory environment
4.37
4.75
Presence of skilled labour
3.32
3.75
Cold climate
3.11
4.25
Good internet connectivity
4.32
4.38
Easy access to substantial electricity supply
4.37
4.88
Low electricity cost
4.47
4.88
Cheap land
3.58
3.75
Special incentives for mining-related activities
3.95
4.13
Low crime rate
3.63
3.38
Table 9: Miners rank major concerns over additional risks
Concerns
Small Miners
(incl. Individuals)
Large Miners
2017
2018
2017
2018
Centralisation of hashpower in a particular geographic area
(location)
3.70
3.89
3.11
3.69
Centralisation of hashpower in the hands of a few (control)
3.89
4.41
3.30
4.00
Centralisation of mining equipment production in a particular
geographic area
3.35
3.70
2.10
3.50
Risk of state-sponsored attack on a cryptoasset system
N/A
3.37
N/A
2.92
Unfavourable global regulation related to cryptoassets
N/A
3.25
N/A
3.15
Unfavourable global regulation related to cryptoasset mining
N/A
3.33
N/A
3.00
Criminal use of cryptoassets
N/A
3.20
N/A
2.77
Popularity of pre-mined/'mining-less' cryptoassets
N/A
3.07
N/A
2.77
Too many cryptoassets in the market
N/A
3.11
N/A
2.08
Respondents scored these categories on a 1-5 scale:
1: Not important at all 2: Not important 3: Neutral 4: Somewhat important
5: Very important
Lowest average score
Highest average score
Respondents scored these categories on a 1-5 scale:
1: Not concerned at all 2: Not concerned 3: Neutral 4: Somewhat concerned 5: Very concerned
Lowest average score
Highest average score
Cambridge Centre for Alternative Finance
10 Trumpington Street
Cambridge CB2 1QA
United Kingdom
E: ccaf@jbs.cam.ac.uk
T: +44 (0)1223 339111