Proof-of-Work vs.

Proof-of-Stake
Securing the chain

August 2022
 Table of contents
1. Introduction 3
2. The anatomy of blockchain protocols 4
3. Trade-offs 12
4. Conclusion and outlook 31

Disclosures
This report has been prepared solely for informative purposes and should not be the basis for making investment decisions
or be construed as a recommendation to engage in investment transactions or be taken to suggest an investment strategy
with respect to any financial instrument or the issuers thereof. This report has not been prepared in accordance with the legal
requirements designed to promote the independence of investment research and is not subject to any prohibition on dealing
ahead of the dissemination of investment research. Reports issued by Payward, Inc. (“Kraken”) or its affiliates are not related to
the provision of advisory services regarding investment, tax, legal, financial, accounting, consulting or any other related services
and are not recommendations to buy, sell, or hold any asset. The information contained in this report is based on sources
considered to be reliable, but not guaranteed to be accurate or complete. Any opinions or estimates expressed herein reflect a
judgment made as of this date, and are subject to change without notice. Kraken will not be liable whatsoever for any direct or
consequential loss arising from the use of this publication/communication or its contents. Kraken and its affiliates hold positions
in digital assets and may now or in the future hold a position in the subject of this research.
1.
Introduction

At the heart of every blockchain lies a protocol that defines how pseudonymous
individuals reach consensus in a globally distributed network. Arguably the most critical
component of a blockchain’s consensus method is its Sybil resistance mechanism, such
as Proof-of-Work (PoW) or Proof-of-Stake (PoS). As its name implies, Sybil resistance
mechanisms protect blockchain networks against Sybil attacks, including spam nodes
and 51% attacks. These mechanisms also regulate the selection of a block author and
incentivize network nodes to behave honestly.

A rivalry exists between PoW and PoS among crypto communities that surfaces key
questions of network security, sustainability, barriers to entry, and decentralization.
Though many claim absolutely which Sybil resistance mechanism is best for blockchain
networks, the reality is not black and white. Neither Sybil resistance mechanism is perfect.
Instead, it is essential to understand the trade-offs between each before concluding on the
optimal choice for a particular blockchain network.

This report examines the nuances behind consensus methods and their Sybil resistance
mechanisms, describes trade-offs between PoW- and PoS-based systems, and compares
key metrics for both mechanisms. Moreover, it makes the case that a given blockchain’s
utility – hard money or smart contracts, for example – determines which characteristics
a network should optimize for. The design choice is often referred to as the “blockchain
trilemma,” which argues that a blockchain network must balance trade-offs between
decentralization, scalability, and security. A highly scalable network, for example, is said
to optimize scalability at the expense of decentralization and network security. These
choices also impact which Sybil resistance mechanism a particular protocol should utilize.
Though other Sybil resistance mechanisms exist, including Proof-of-Authority and Proof-
of-Spacetime, this note solely focuses on PoW and PoS. Currently, PoW and PoS make up
the lion's share of the major Layer-1 (L1) blockchain networks by market capitalization.

July 2022 For more Kraken Intelligence content: kraken.com/subscribe/intelligence Page 3

2.
The anatomy of blockchain protocols

Every blockchain follows an underlying protocol that determines its block selection
process, how network nodes validate transactions, transaction finality characteristics,
supply issuance, supply distribution, and what determines the “true” state of the network.
Furthermore, the blockchain protocol provides an incentive structure for network
participants to behave honestly with each other while discouraging bad actors.

The Sybil resistance mechanism establishes a cost that disincentivizes a bad actor from
maliciously taking over as the sole arbiter of consensus. The consensus method describes
how nodes coordinate to agree on the validity of transactions and the state of the network.
Examples of consensus methods include Nakamoto consensus and Byzantine Fault-Tolerance.

Sybil resistance mechanism

Sybil resistance mechanisms are compatible with various consensus methods, such as
Nakamoto consensus and Practical Byzantine Fault-Tolerance (PBFT), but they are not sole
descriptors for consensus. Instead, these mechanisms establish rules that nodes must follow
to append data to a blockchain. One of their core functions is to deter Sybil attacks.

A Sybil attack is an exploit against an online network whereby a small number of entities (as
few as one) attempt to take control of the whole network by leveraging multiple accounts,
nodes, or computers. On social media platforms, a user can generate multiple accounts and
spam the network, effectively "taking over" the conversation. On a blockchain, bad actors
might run multiple nodes to achieve an overwhelming influence on the network.

July 2022 For more Kraken Intelligence content: kraken.com/subscribe/intelligence Page 4

In Sybil attacks, dishonest nodes try to out-vote honest nodes on the network by creating
enough “Sybil identities.” The word "Sybil" comes from a case study about a woman
named Sybil Dorsett, a pseudonym for Shirley Ardell Mason, who received treatment for
dissociative identity disorder, or multiple personality disorder.1 Once an attacker creates
enough Sybil identities to disproportionately influence a crypto network, they can refuse to
receive or transmit blocks, effectively preventing other users from the network.

The most commonly-known Sybil attack in the crypto space is the "51% attack," where
attackers take over most of the network computing power, commonly referred to as “hash
rate.” In such cases, they may theoretically influence the ordering of transactions, prevent
the confirmation of new transactions, and double-spend their cryptoassets. Resistance to
these attacks is essential for a well-functioning, decentralized blockchain.

PoW and PoS are economic deterrents to Sybil attacks because they require users to expend
energy or lock-up collateral to participate in network validation. The crux of a Sybil
resistance mechanism is that it requires each validator or miner to have "skin-in-the-game"
to participate in a decentralized, cryptographic system. Sybil resistance mechanisms are
also known as “block author selectors” because they designate a validator or miner to add
a block to the chain.2 Readers should note that these mechanisms are not cures against
Sybil attacks; instead, they make it impractical for an attacker to carry out a Sybil attack
successfully by:

a. Encouraging participants to reach a consensus on the state of the blockchain through a

competitive process, such as mining or staking;

b. Punishing bad actors that try to stall the network from reaching a consensus; and

c. Rewarding some or all participants for behaving honestly and coming to a consensus
(e.g., block subsidies, transaction fees).

August 2022 For more Kraken Intelligence content: kraken.com/subscribe/intelligence Page 5

Figure 1
Sybil Resistance Mechanism Overview

Sybil Resistance Market

Mechanism Competition Method Penalty for Misbehavior Dominance

Proof-of-Work Computational Solve mathematical puzzles Proposing an invalid block results in 58%
(PoW) work using computational hardware wasted time, energy, and money

Proof-of-Stake Financial stake Lock up funds in a smart The protocol can destroy a validator's 12%
(PoS) contract stake or bar them from participating
in consensus if they fail to step up
when called upon or sign invalid
blocks

Other* - - - 30%

Non-PoW/PoS Sybil resistance mechanism examples, including but not limited to:

Proof-of-Authority Reputation Validators undergo The protocol can exclude nodes that
(PoA) authentication to participate cheat or go offline from consensus
and the consortium of approved
validators can impose other penalties

Proof-of-Space Disk space Solve mathematical puzzles by The protocol can destroy a validator's
(PoSp) dedicating disk space stake or bar them from participating
in consensus if they fail to step up
when called upon or sign invalid
blocks

Proof-of-Elapsed Time Fair lottery Each node must wait for a Since PoET is designed for
(PoET) randomly chosen period; permissioned blockchains, the
the first to complete the protocol's leaders can boot any
designated waiting time wins misbehaving nodes from the network
the new block

Proof-of-Burn Coin burn Burn coins to win the right to Proposing an invalid block results in
(PoB) propose a block wasted time and money

Source: Kraken Intelligence, CoinGecko

*Note: The market dominance figure for “other” includes tokens, which are cryptoassets built on top of existing L1 blockchains. Tokens do not have
their own native blockchain and must follow the protocol rules of the chain they operate on.

August 2022 For more Kraken Intelligence content: kraken.com/subscribe/intelligence Page 6

Consensus method

Blockchain consensus methods mitigate the challenge of achieving consensus in a globally

distributed, digital world by enabling users to validate entries into the blockchain ledger,
help synchronize data, and bolster the network’s security. Such methods must ensure that
all network participants can agree on a single source of "truth," even if some nodes fail;
put differently, they must be Byzantine Fault Tolerant (BFT). The concept of BFT derives
from the Byzantine Generals' Problem, a game theory problem that describes the difficulty
decentralized parties have in achieving consensus without relying on a trusted central
party. Initially conceived in 1982 as a logical dilemma, a group of Byzantine generals
must perfectly coordinate an attack with the added challenge that they cannot directly
communicate with each other.

A Byzantine army is besieging an enemy city and has the territory surrounded. The army
splits into several divisions, each commanded by a different general. After observing the
enemy, they must agree on a joint action plan. The generals can win if they all attack
simultaneously but will lose if they strike at different times. However, the generals can
only communicate with each other by messenger. Some of the generals may be traitors, or
any messages sent or received could have been intercepted or deceptively sent by the enemy
to prevent the honest generals from reaching a consensus. Thus, the Byzantine Generals
Problem highlights a common problem among distributed networks: can independent
participants of a distributed network form an agreement?3

In other words, the consensus method allows network participants to propose and
confirm transactions and agree on the state of the distributed ledger in near real-time.
PoW and PoS choose a block author and defend against Sybil attacks but do not, in and
of themselves, describe a blockchain’s consensus method. Instead, a consensus method
combines a Sybil resistance mechanism with a chain selection rule, which decides which is
the valid version of the blockchain.

August 2022 For more Kraken Intelligence content: kraken.com/subscribe/intelligence Page 7

Many L1 blockchains, including Bitcoin and Cardano, use the “longest chain” rule, known
as Nakamoto-style consensus, meaning that nodes will accept whichever blockchain is
the longest as the true state of the transaction ledger. For PoW chains, the chain’s total
cumulative PoW determines the longest chain. Conversely, PoS chains define the true
state of the network as the chain with the most votes. Another common chain selection
rule used by L1 blockchains, including Ethereum and Conflux, is the “heaviest chain rule,”
a variation of the longest-chain rule that includes orphan blocks or blocks that were
previously authored within the blockchain network but were not accepted at the time. This
rule allows a peer-to-peer network of distributed nodes to achieve BFT because each node
unambiguously agrees to follow a source of truth derived from a verifiable set of records.

Before Bitcoin, distributed cryptographic systems could only achieve up to 33% fault
tolerance utilizing a BFT-style consensus method. Transactions achieve finality in BFT-
style, distributed networks when 66% of the aggregate financial stake in the network
reaches an agreement. In other words, anyone that can accumulate more than 33% of the
total value staked on the network can prevent users from finalizing transactions, reaching
a consensus, and censor users.

Depending on the consensus method of the network, this could result in the network
ceasing to produce blocks until it gets 66% agreement on the block or continuing to
produce blocks but not reaching a final agreement on the content of the blocks. Satoshi’s
invention of Nakamoto-style consensus incentivized network participants to act honestly
in the absence of trust and increased the theoretical fault tolerance of distributed systems
from 33% to 50%, effectively deterring Sybil attacks that could lead to double-spending.

Transactions finality is always “probabilistic” on Nakamoto-style consensus protocols

because nodes come to a consensus around the longest chain. The node that proposes
the transactions that go into a particular block is not predetermined like in BFT-style
protocols, meaning it is never 100% guaranteed that a transaction is irreversible as a longer
chain could exist. It is considered “probabilistic” because the probability of a transaction
reversal decreases as a chain gets longer, providing near certainty in the transaction’s
irreversibility after a certain period has passed. For instance, most network participants
typically wait for 40-60 confirmations for probabilistic finality on Dogecoin and six

August 2022 For more Kraken Intelligence content: kraken.com/subscribe/intelligence Page 8

confirmations for Bitcoin. Confirmations on a blockchain refer to the validation of blocks,
meaning that a transaction receives another confirmation every time miners add a block
to the chain. In these networks, this amount of confirmations makes it improbable to
reorganize the blockchain, but it is never theoretically impossible.

On the other hand, transactions on BFT-style networks are “deterministic” because rules
determine who can vote on a transaction and exactly how many votes are needed before
everyone can agree the transaction is 100% final. Once a transaction achieves finality, there
is no way a longer chain can exist, as there is no uncertainty in the process.

The success of Bitcoin and its pioneering PoW-based Nakamoto consensus method showed
that decentralized networks could work and accrue value for users and network stakeholders
by solving fundamental problems such as the Byzantine Generals Problem, Sybil resistance,
and economic incentivization. Since then, thousands of blockchain networks have launched
using similar mechanisms to deter Sybil attacks and achieve a consensus.

Figure 2
Blockchain Consensus Method Examples
Sybil
Consensus Method Resistance Block Information Chain Selection Transaction Incentive
(Examples) Mechanism Validation Propagation Rule Finality Structure Fault-Tolerance

Nakamoto (Bitcoin, PoW PoW Gossip Longest-chain Probabilistic Block subsidy 50%
Litecoin) verification and transaction computational
fees power

Nakamoto-style, PoW (Ethash) PoW Gossip Heaviest-chain Probabilistic Block subsidy 50%
GHOST (Ethereum) verification and transaction computational
fees power

Nakamoto-style, PoS PoS Gossip Longest-chain Probabilistic Block subsidy 50% financial
chain-based PoS verification and transaction stake
(Peercoin, Nxt, fees
PoAct)

Nakamoto-style, PoS-based Proposer Broadcast Longest-chain Probabilistic Block subsidy 50% financial
committee-based committee eligbility among stake
PoS (Ouroboros, election verification committee
Praos, CoA, Snow
White)

BFT-style PoS-based Proposer Broadcast BFT (adapted Deterministic Block subsidy 33% financial
PoS—Algorand committee eligibility among Byzantine stake
(Algorand) election verification committee agreement)

BFT-style PoS— PoS-based Proposer Broadcast BFT (adapted Deterministic Block subsidy 33% financial
Tendermint round robin eligibility among distributed stake
(Cosmos Hub) verification committee ledger system)

Source: Kraken Intelligence, Protocol whitepapers

August 2022 For more Kraken Intelligence content: kraken.com/subscribe/intelligence Page 9

At a high level, some popular blockchains utilize the following consensus methods:

• Cardano—Ouroboros, the blockchain protocol implemented on Cardano, utilizes

Nakamoto-style consensus similar to Bitcoin, where nodes follow the "longest
chain rule." However, unlike Bitcoin's PoW mechanism that uses hash power or
energy to create a new block, Ouroboros implements PoS, which uses the network's
native coin (i.e., ADA) to influence how often the block selection mechanism chooses
a validator node to create a new block. Cardano's use of Nakamoto consensus
and PoS differentiates it from other major 3rd-generation protocols, which often
combine PoS with BFT-style consensus protocols. Instead of following the longest-
chain rule, BFT-style consensus protocols come to a consensus in a quorum vote,
where a 2⁄3 majority is required to confirm a block.
• Ethereum—The Ethereum blockchain's consensus method also uses Nakamoto-
style consensus combined with Ethash, the blockchain’s modified version of PoW.
The Ethash PoW algorithm depends on generating and analyzing a large, frequently
accessed dataset, known as a directed acyclic graph (DAG). Ethereum’s chain
selection rule is known as Greedy Heaviest Observed Subtree (GHOST), or the
“heaviest chain rule.”5
• Ethereum 2.0—The ETH 2.0 Beacon Chain uses a PoS-based consensus method
called Casper the Friendly Finality Gadget.6 Casper is a partial consensus method
combining PoS and BFT-style consensus. The method inherited its core design from
the PBFT consensus method while adding new mechanisms and simplifying several
rules. This consensus method is similar to Nakamoto Consensus in that it defines
the "true" chain as the chain with the most attestations.
• Solana—Tower BFT is the consensus method Solana utilizes on top of the innovative
Proof-of-History (PoH) timing mechanism in conjunction with PoS. Because nodes
in a distributed network cannot trust the timestamp on messages received from
other nodes, distributed networks cannot form a consensus on the time and order
in which events happen. Solana uses PoH to overcome this problem by establishing
a cryptographically safe source of time throughout the network. PoH is a sequence
of computations that provides a digital record to prove that an event occurred on

August 2022 For more Kraken Intelligence content: kraken.com/subscribe/intelligence Page 10

 the network at any given time. At a high level, Tower BFT works in that when a node
votes on a specific fork, they agree to lock themselves out of voting on an opposing
fork for a period. As they continue to vote on the same fork, the time they are locked
out rises exponentially until they reach a maximum lockout of 32 votes for the same
fork. When nodes on the network hit this maximum vote lockout, they will earn
inflation incentives.7

Though many blockchains have developed unique methods of achieving consensus and
BFT, readers should carefully note the difference between a Sybil resistance mechanism
and a consensus method. This understanding is critical in determining whether PoW or
PoS is a superior model for blockchain design. Sybil resistance mechanisms work with a
chain selection rule to form the consensus method that supports a functional blockchain.
Furthermore, these Sybil resistance mechanisms have significant trade-offs that result in
different desired outcomes for the functionality of a blockchain.

August 2022 For more Kraken Intelligence content: kraken.com/subscribe/intelligence Page 11

3.
Trade-offs

Proof-of-work (PoW): bridging the physical to the digital

One of the most popular Sybil resistance mechanisms used in crypto networks is PoW, a
cryptographic proof in which one party (the miner) proves to other network participants
(nodes) that they have solved a sufficiently complex problem that requires computational
effort. Introduced in the 1990s to mitigate email spam, PoW involves the exertion of
computational power to solve a moderately difficult and random "puzzle" to gain access to
the resource, thus preventing frivolous use.8,9 At the time, the goal of PoW was to require
computers to perform a small amount of "work" before sending an email. This work would
require trivial computing power for someone sending a legitimate email while creating high
computational costs for those sending mass emails.

In PoW blockchains, miners must gather information and try to guess a solution to a
cryptographic puzzle. Once miners solve the puzzle, they share their results with other
nodes to verify their "work" before adding the block to the chain. Because miners must
consume energy in the PoW process and nodes can easily verify any block's validity,
malicious miners who try to add an invalid block waste time, energy, and resources.
Conversely, the protocol rewards honest miners that successfully mine a valid block.

This mechanism incentivizes honest nodes to act in good faith while discouraging bad
actors on the network. In short, PoW networks consume energy to maintain the continued
functionality of the distributed blockchain ledger and fairly distribute the cryptoasset's
supply without a centralized overseer. This PoW competition among miners comes with
several benefits and considerations.

July 2022 For more Kraken Intelligence content: kraken.com/subscribe/intelligence Page 12

PoW Benefits
• PoW blockchains are battle-tested at scale since crypto protocols implemented the
Sybil resistance mechanism throughout most of crypto history, securing billions
of dollars worth of value for nearly a decade. Still, readers should note that PoW
blockchains are not automatically impervious to hacks as smaller PoW blockchains,
including Vertcoin, Verge, and Ethereum Classic, have experienced several successful
attacks.
• Carrying out a 51% attack on an established PoW blockchain network, such as
Bitcoin, Ethereum, Litecoin, or Dogecoin, requires so much computing power that
it is meaningfully expensive; any attacker would spend more than they could earn
to attack the network. This financial cost disincentivizes bad actors from behaving
maliciously on the network and secures the blockchain.

Figure 3
Theoretical Requirements for 51% Attack on PoW Blockchains Using Bitmain ASICs

ASIC Model Bitmain Antminer S19 Antminer L7 Antminer Z15 Antminer Antminer
D7 DR3
Hashing SHA-256 Scrypt Equihash X11 Blake256r14
Algorithm
Individual Equipment 95 TH/s 0.00905 TH/s 0.0000004 TH/s 1.286 TH/s 7.580 TH/s
Mining Hashrate
Equipment Equipment $3,990 $13,999 $7,999 $1,778 $1,999
Specifications Cost
Power 3,250W 3,260W 1,510W 3,148W 1,410W
Consumption
Electricity $9.18 $9.21 $4.27 $8.89 $3.98
Cost/Day*

Cryptoasset BTC BCH BSV XEC LTC DOGE ZEC DASH DCR

Market $384.8B $2.2B $1.2B $715.4M $3.7B $8.3B $816.4M $496.6M $325.5M
Cryptoasset Capitalization
Specifications Hashrate 205 1.15 0.65 0.32 379 TH/s 329 TH/s 0.00897 TH/s 3,107 TH/s 95,178 TH/s
EH/s EH/s EH/s EH/s
Hashrate 98.93% 0.55% 0.31% 0.15% 0.00018% 0.00016% 0.000000004% 0.0015% 0.046%
(%share)
Hashrate 207.2 1.16 0.66 0.323 382 TH/s 333 TH/s 0.00906 TH/s 3,138 TH/s 96,130 TH/s
EH/s EH/s EH/s EH/s

ASICs 2.16M 12.1K 6.8K 3.4K 42.3K 36.8K 21.6K 2.4K 12.7K
51% Attack
Requirements Hardware $8.6B $48.3M $27.3M $13.6M $591.7M $514.5M $172.5M $4.3M $25.4M
Cost
Electricity $19.8M $111.1K $62.8K $31.2K $389.2K $338.5K $92.0K $21.7K $50.5K
Cost/Day*

Source: Kraken Intelligence, Bitmain, Newegg

*Note: Energy costs assume a fixed cost of 11.77 cents per kWh, the average electric price a business customer in the U.S. pays for electricity as of
June 2022.10

August 2022 For more Kraken Intelligence content: kraken.com/subscribe/intelligence Page 13

 • Large intermediaries may have less influence on the governance of PoW-based
cryptoassets, as evidenced by the failure of SegWit2x on Bitcoin in 2017.11 Though most
of the large exchanges and custodians supported the movement at the time, nodes
did not implement the software upgrade due to a lack of network-wide consensus.12
Should a similar situation emerge today on one of the larger PoS networks, custodians
with large amounts of coin supply may influence protocol governance.
• PoW incentivizes miner operations to distribute geographically and organizationally,
decentralizing PoW cryptoassets as a whole. This incentive stems from the fact
that electricity, miners’ highest variable cost, varies in price depending on location.
Because miners constantly pursue lower energy costs and cheap energy exists in
remote locations across the globe, mining operators also distribute their operations
worldwide. Bitcoin, which houses most of the hash power of all PoW blockchains, is a
fine example of this incentive structure. Figure 4 shows that this incentive structure
caused a geographic centralization of mining on the network in its early years as
China was abundant with cheap electricity. Nonetheless, miners have increasingly
distributed internationally in recent years in search of cheaper options. Critics
may argue that this decentralization of hash power was solely due to China’s ban
on mining in May 2021; however, data suggests the distribution began long before
the crackdown. From September 2019 to May 2021, China’s hash power dominance
dropped from 76% to 44%. Since the ban, that figure has dipped further to 21%.

August 2022 For more Kraken Intelligence content: kraken.com/subscribe/intelligence Page 14

Figure 4
Distribution of Bitcoin Mining Hash Power by Country

Source: Kraken Intelligence, Cambridge Centre for Alternative Finance

• Incentive structure includes measures to prevent constant forking, while forking

is not automatically discouraged by PoS systems. The “nothing at stake” problem
occurs when a validator signs off on both sides of a fork to earn rewards on
both blockchains, afflicting PoS protocols. In a PoS fork, pre-fork holders are
presented with a financial incentive to stake the same amount of coins on both
the original blockchain and the forked blockchain to increase their total returns.
PoW networks like Bitcoin correlate security with hash power, so a fork would not
necessarily maintain the hash power of the original network, and thus neither
the security. This feature provides a high cost to forking on PoW systems that
strongly discourages such an endeavor. Critics may argue that the existence of

August 2022 For more Kraken Intelligence content: kraken.com/subscribe/intelligence Page 15

 merge mining, the process of mining two blockchains that use the same hashing
algorithm at the same time, introduces the equivalent problem on PoW as nothing-
at-stake does for PoS.13 However, because a new PoW blockchain created from a
hard fork would need to retain the same algorithm as the original blockchain for
merge mining capabilities, the issue does not always arise. For instance, Bitcoin
Gold and Ethereum Classic did not retain their respective SHA-256 and Ethash
hashing algorithms following their respective 2017 hard forks.14,15 Moreover, PoW
blockchain forks that retain the same hashing algorithm as the original blockchain
typically fall behind significantly in terms of hash power, as evidenced by historic
blockchain forks such as Bitcoin Cash and Bitcoin SV. While Bitcoin Cash is one of
the most successful Bitcoin hard forks ever, its hashrate of 1.12 EH/s falls -99.45%
behind the 204.6 EH/s on its original chain, Bitcoin, and Bitcoin SV’s hashrate of
0.66 EH/s is -41.1% behind Bitcoin Cash.16
• It is debatably harder to perform bribery attacks on PoW than on PoS due to
the nothing-at-stake problem, which gives PoS validators a greater incentive to
behave dishonestly. Bribery attacks rely on a bad actor successfully bribing enough
validators or miners to work on specific blocks or forks so the colluding miner(s)
or validator(s) can present arbitrary transactions as valid and have dishonest nodes
paid to verify them. In a bribery attack, a bad actor sends a transaction, discretely
builds an alternative chain based on the prior block until the transaction receives
enough confirmations and the attacker’s chain is longer than the valid chain,
and then publishes the invalid chain as the new valid blockchain to reverse the
transaction.17,18 In a PoW system, a similar attack would require the attacker to
bribe the majority of miners by hash rate. Since miners lose resources spent on
computations if the attack fails, there are strong assumptions that the number
of bribes is prohibitively high. Though bribery attacks are theoretically harder to
conduct on an established PoW network, it is also impractical to conduct on a PoS
network large enough to make it extremely expensive to attain enough validator
votes for such an attack.

August 2022 For more Kraken Intelligence content: kraken.com/subscribe/intelligence Page 16

PoW Considerations
• Requires energy consumption to achieve Sybil resistance. This feature has spurred a
debate on the carbon footprint of cryptocurrency networks.
• Smaller PoW blockchains with low hash power have low security and thus are prone
to 51% attacks. Some larger PoW projects that have experienced 51% attacks in the past
include Ethereum Classic, Verge, Bitcoin Gold, and Vertcoin.19,20,21,22,23
• Some argue that it is difficult for individual miners to continuously upgrade their
hardware to compete effectively, eventually leading to a centralization of mining
towards corporate players. Data shows that the top three mining pools in Bitcoin
control over 52% of the current hash rate, suggesting a group effort among these
organizations could theoretically attack the network successfully.24 The pool selects
transactions to put in the blocks that everyone in the pool is working on, receives
the pool’s mining rewards, and holds the assets until individual hardware operators
withdraw them. However, it is worth noting that individual miners could stop
contributing power to pools that show any sign of wrongdoing since it is not in their
best interest, suggesting arguments for the dangers of mining pool centralization are
theoretically unsound. Moreover, mining pool operators have a long-term stake in
the Bitcoin network and little incentive to attempt an attack as they have a financial
incentive to behave honestly. However, this is equally true in PoS networks as
validators must hold the blockchain’s native cryptoasset to stake.
• Since nodes can operate anonymously, blocking a malicious miner from participating
in the network is impossible, and there is no way to confiscate their mining
equipment for misbehaving. Bad actors on PoW networks only waste time, energy,
and the funds used to attempt the attack, while the same attack on PoS can cause
hackers to lose their stake and bar them from the validation process.
• PoW systems that follow Nakamoto-style consensus could encounter “selfish mining
attacks,” a deceitful endeavor where a miner or group of miners finds a new block and
withholds it from the blockchain. First identified by Cornell researchers Emin Gün
Sirer and Ittay Eyal in a 2013 paper, selfish mining attacks create a blockchain fork,

August 2022 For more Kraken Intelligence content: kraken.com/subscribe/intelligence Page 17

 which the malicious miner(s) work on to get ahead of the original blockchain.25 If
the blockchain fork becomes longer than the original blockchain, the miner can
introduce its newest block to the network and effectively overwrite the original
blockchain with the blockchain fork. This attack alters the blockchain to allow the
malicious miners to steal cryptoassets from other users or double-spend funds. Still,
selfish mining attacks are theoretical as they have not occurred on a live blockchain.
• Some critics contend that PoW’s trial-and-error architecture naturally entails a
delay in block production, meaning that fees rise in times of congestion. However,
this critique fails to understand what drives throughput. Block size and the number
of bytes (and hence, transactions) that can fit into a block, primarily determines
throughput, not the time between blocks. For example, a blockchain designed to
produce one block per second with 1,000 transactions per block could have the
same throughput as a blockchain that produces one block per minute that is large
enough to fit 60,000 transactions. Moreover, the same critics may argue that
material blockchain fees are unhealthy for a scalable network; instead, cryptoassets
with zero transaction fees are preferable. Having fees is arguably healthy for a
public blockchain system because it eliminates the spam and DDoS problem by
making it costly to insert junk data. Fees also promote a competitive environment
among validators, making it prohibitively expensive for single parties to attack
a network successfully. Spam and DDoS attacks have historically plagued zero-
and low-fee networks like Nano and Solana due to low barriers to entry when
conducting a transaction, causing network nodes to lose synchronization.26,27 These
spam attacks have caused user transactions to fail while significantly decreasing
overall transaction throughput. Furthermore, such events could cause users to lose
confidence in projects with said vulnerabilities, causing an outflow of users and
negative selling pressure on the cryptoasset.

August 2022 For more Kraken Intelligence content: kraken.com/subscribe/intelligence Page 18

Proof-of-stake (PoS): earning in a digital world

In PoS systems, the native coin stores value and voting power rather than just value as in
PoW systems. Peercoin first implemented PoS in 2012 to create a blockchain network that
achieved Sybil resistance in a BFT way while consuming a fraction of the energy.28 Rather
than relying on computers racing to generate the appropriate hash, the act of locking up
coins, or staking, determines participation in a PoS protocol. This mechanism attempts
to reduce the computational cost of PoW schemes by selecting validators in proportion
to their quantity of staked holdings in the associated cryptoasset. Using a set of factors
determined by the protocol, the PoS mechanism pseudo-randomly selects a validator node
actively staking to propose the next block to the blockchain. When the mechanism elects a
validator node, the node must verify the validity of the transactions within the block, sign
it, and propose the block to the network for further validation. PoS-based blockchains come
with notable benefits and considerations that differ from PoW.

PoS Benefits
• Achieving Sybil resistance requires virtually no energy consumption. For reference,
Ethereum developers estimate that its transition to a PoS system will reduce energy
consumption by more than 99.9%.29 Moreover, thanks to the low energy requirement,
less native coin issuance is required to incentivize participation in the validation process.
• PoS eliminates the need for validators to purchase and upgrade hardware continuously.
• Honest validators could decide to forcibly remove attackers from the network and
destroy their staked cryptoassets, providing economic defenses against a 51% attack.
For example, the Ethereum 2.0 protocol includes a “correlation penalty” where
validators forfeit ETH rewards if they fail to participate when called upon, and the
protocol destroys, or “slashes,” their existing stake if they propose multiple blocks
in a single slot or submit contradictory votes. The amount of ETH slashed depends
on how many dishonest validators the protocol slashes around the same time. This
penalty can result in the slashing of roughly 1% of a validator’s stake if they are

August 2022 For more Kraken Intelligence content: kraken.com/subscribe/intelligence Page 19

 penalized on their own or 100% of the validator’s stake during a mass slashing event.
The protocol imposes the penalty halfway through a forced exit period that begins
with an immediate penalty (up to Ξ0.5) on day 1, the correlation penalty on day 18,
and finally, ejection from the network on day 36. Furthermore, the dishonest node
receives minor attestation penalties daily because they are present on the network
without submitting votes.30 However while PoS protocols such as Ethereum 2.0,
Polkadot, Solana, and Cosmos, enforce a “slashing” penalty, some PoS protocols,
including Cardano, Avalanche, and Algorand, do not include a slashing feature.
• The time it takes for PoS blockchains to choose a validator is faster and has less block
variance than PoW mining competition, allowing for increased efficiency.
• Lower barrier to entry since the protocol only requires funds to participate in the
block validation process, rather than warehousing, energy contracts, and specialized
hardware as is required in PoW.

PoS Considerations
• Not as extensively tested as PoW, which has secured billions of dollars worth of value
for nearly a decade. Particular implementations of PoS could introduce black swan
attack vectors, decreasing the overall security of the blockchain.
• The initial supply distribution in PoS systems can introduce voter concentration
depending on the free market accessibility to initial supply distributions.
• PoS blockchains using BFT-style consensus reduce fault tolerance from 50% to 33%,
as Nakamoto-style consensus is critical in resisting 51% attacks. Suppose a validator
or a coordinated set of validators own more than 33% of the financial stake in these
blockchains. In that case, they can attack the network by reverting transactions,
censoring transactions, or stopping network participants from reaching a consensus.

August 2022 For more Kraken Intelligence content: kraken.com/subscribe/intelligence Page 20

 • PoS is potentially more vulnerable to centralization than PoW because capital
ownership determines network control, which is more centralized than labor and
cheap energy. In a BFT-style PoS network worth $100 billion, where users stake
10% of tokens, any party able to allocate more than $33 billion (>33%) can take over
the network by locking their assets in a staking contract. In a PoW network using
Nakamoto-style consensus, attacks require most of the mining equipment and
labor. Attacking a network with $10 billion of security would require acquiring
specialized hardware, space, and energy contracts to mine at a larger scale than
the entire network and deploy the labor to execute the attack. If such an attack
were underway, the entire network would likely become aware of the immense
demand for mining equipment and electricity ahead of time. Because carrying
out a 33% attack on a PoS network requires holding 33% of staked tokens, those
with large amounts of coins, such as early adopters, exchanges, and custodians,
can overwhelmingly influence the rules of the network. They can also accumulate
more of the coin via staking, causing a positive feedback loop that can increase
centralization. There are already examples of exchanges used to influence PoS
networks, including when Tron founder Justin Sun worked with several exchanges
to obtain influence on the Steem network by voting with their user funds in favor
of his proposal.31 The figure below analyzes some major blockchain’s and the
minimum number of validators required for a 33% attack within those networks,
suggesting that the requirements to halt PoS-based networks such as Solana,
Algorand, Avalanche, and Cosmos Hub, are costly but only require collaboration
among a small set of validators.

August 2022 For more Kraken Intelligence content: kraken.com/subscribe/intelligence Page 21

Figure 5
Minimum Number of Validators Required for a 33% Attack
Solana Algorand Avalanche Cosmos Hub

Validator Stake ($)/ Cumulative Stake ($)/ Cumulative Stake/ Stake ($)/ Cumulative Stake ($)/ Cumulative
Share (%) Stake/Share Share (%) Share Share (%) Stake/Share Share (%) Stake/Share
1 $322M (2.56%) $322M (2.56%) $32.5M (3.15%) $32.5M (3.15%) $49.0M (1.22%) $49.0M (1.22%) $84M (6.30%) $84M (6.30%)

2 $266M (2.11%) $589M (4.67%) $26.8M $59.3M $49.0M (1.22%) $98.0M $77M $161M
(2.61%) (5.76%) (2.45%) (5.73%) (12.03%)
3 $262M (2.08%) $851M (6.75%) $26.4M $85.7M $49.0M (1.22%) $147.0M $76M $237M
(2.56%) (8.32%) (3.67%) (5.63%) (17.66%)
4 $232M (1.84%) $1.1B $26.2M $111.9M $49.0M (1.22%) $196.0M $72M $299M
(8.58%) (2.54%) (10.86%) (4.90%) (4.60%) (22.26%)
5 $196M (1.56%) $1.3B $26.2M $138.0M $49.0M (1.22%) $245.0M $61M $360M
(10.14%) (2.54%) (13.40%) (6.12%) (4.54%) (26.80%)
6 $196M (1.56%) $1.5B $23.3M $161.3M $48.7M $293.7M $57M $416M
(11.695) (2.27%) (15.67%) (1.22%) (7.34%) (4.23%) (31.03%)
7 $178M (1.41%) $1.7B $23.3M $184.6M $48.6M $342.3M $48M $464M
(13.10%) (2.26%) (17.93%) (1.22%) (8.56%) (3.57%) (34.60%)
8 $177M (1.41%) $1.8B $21.9M $206.5M $48.1M $390.5M $47M $511M
(14.51%) (2.13%) (20.05%) (1.20%) (9.76%) (3.47%) (38.07%)
9 $169M (1.34%) $2.0B $21.9M $228.4M $48.1M $438.6M $47M $557M
(15.85%) (2.13%) (22.18%) (1.20%) (10.96%) (3.47%) (41.54%)
10 $166M (1.32%) $2.2B $17.6M $246.0M $46.8M $485.4M $44M $602M
(17.17%) (1.71%) (23.89%) (1.17%) (12.13%) (3.49%) (44.83%)
11 $149M (1.18%) $2.3B $16.2M $262.2M $46.6M $532.0M $40M $642M
(18.35%) (1.58%) (25.46%) (1.17%) (1.30%) (3.01%) (47.85%)
12 $149M (1.18%) $2.5B $15.6M $277.8M $46.2M $578.3M $40M $682M
(19.53%) (1.52%) (26.98%) (1.16%) (14.46%) (2.99%) (50.84%)
13 $147M (1.17%) $2.6B $15.4M $293.3M $45.0M $623.3M $34M $716M
(20.70%) (1.50%) (28.48%) (1.12%) (15.58%) (2.52%) (53.36%)
14 $144M (1.14%) $2.8B $14.9M $308.2M $43.7M $666.9M $29M $745M
(21.84%) (1.45%) (29.93%) (1.09%) (16.67%) (2.16%) (55.52%)
15 $141M (1.12%) $2.9B $14.9M $323.1M $43.1M $710.0M $26M $771M
(22.96%) (1.44%) (31.37%) (1.08%) (17.75%) (1.96%) (57.48%)
16 $135M (1.07%) $3.0B $14.1M $337.1M $42.5M $752.6M $26M $797M
(24.03%) (1.36%) (32.74%) (1.06%) (18.81%) (1.91%) (59.39%)
17 $132M (1.05%) $3.2B $14.0M $351.1M $42.5M $795.1M $21M $818M
(25.07%) (1.36%) (34.09%) (1.06%) (19.87%) (1.60%) (60.98%)
18 $126M (1.00%) $3.3B $13.9M $365.0M $42.3M $837.4M $18M $836M
(26.07%) (1.35%) (35.44%) (1.06%) (20.93%) (1.32%) (62.30%)
19 $115M (0.91%) $3.4B $13.9M $378.8M $42.2M $879.7M $17M $853M
(26.99%) (1.34%) (36.79%) (1.06%) (21.99%) (1.24%) (63.54%)
20 $108M (0.86%) $3.5B $13.7M $392.5M $41.4M $921.1M $15M $867M
(27.84%) (1.33%) (38.12%) (1.03%) (23.02%) (1.08%) (64.62%)
21 $105M (0.83%) $3.6B $13.6M $406.1M $40.4M $961.5M $14M $882M
(28.67%) (1.32%) (39.44%) (1.01%) (24.03%) (1.08%) (65.70%)
22 $104M (0.83%) $3.7B $13.5M $419.6M $39.7M $1.0B $14M $896M
(29.50%) (1.31%) (40.75%) (0.99%) (25.03%) (1.07%) (66.77%)
23 $102M (0.81%) $3.8B $13.5M $433.1M $39.5M $1.0B $14M $910M
(30.31%) (1.31%) (42.06%) (0.99%) (26.02%) (1.02%) (67.79%)
24 $102M (0.81%) $3.9B $13.3M $446.4M $39.5M $1.1B $13M $923M
(31.12%) (1.29%) (43.35%) (0.99%) (27.00%) (1.00%) (68.79%)
25 $99M (0.78%) $4.0B $13.1M $459.5M $39.5M $1.1B $13M $936M
(31.90%) (1.28%) (44.62%) (0.99%) (27.99%) (0.98%) (69.78%)
26 $97M (0.77%) $4.1B $13.1M $472.6M $39.2M $1.2B $13M $949M
(32.67%) (1.27%) (45.89%) (0.98%) (28.97%) (0.97%) (70.74%)
27 $96M (0.76%) $4.2B $13.1M $485.6M $38.8M $1.2B $13M $962M
(33.44%) (1.27%) (47.16%) (0.97%) (29.94%) (0.94%) (71.68%)
28 $94M (0.75%) $4.3B $13.0M $498.6M $38.6M $1.2B $12M $974M
(34.18%) (1.26%) (48.42%) (0.96%) (30.91%) (0.91%) (72.59%)
29 $93M (0.74%) $4.4B $12.9M $511.5M $37.7M $1.3B $11M $985M
(34.92%) (1.25%) (49.67%) (0.94%) (31.85%) (0.85%) (73.44%)
30 $93M (0.73%) $4.5B $12.7M $524.3M $36.9M $1.3B $11M $996M
(35.66%) (1.24%) (50.91%) (0.92%) (32.77%) (0.81%) (74.25%)
31 $93M (0.73%) $4.6B $11.8M $536.0M $36.5M $1.3B $11M $1B
(36.39%) (1.14%) (52.05) (0.91%) (33.68%) (0.80%) (75.04%)

Source: Kraken Intelligence, Block Explorers, CoinGecko

August 2022 For more Kraken Intelligence content: kraken.com/subscribe/intelligence Page 22

 • Stakes can turn rogue and validate incorrect transactions. However, some
protocols have implemented incentive mechanisms to deter this from happening.
For instance, Ethereum, as part of their planned transition to PoS, designed the
“Casper” protocol where such rogue validators are punished by confiscating their
staked cryptocurrencies and barring them from staking again.
• Some PoS protocols may require a large initial investment of the native token to
qualify as a validator, which depends on the size of the network. Thus, PoS network
design may impact centralization depending on how costly it is to partake in
governance. For example, partaking in Ethereum 2.0 governance without a third
party requires an individual to purchase and stake a minimum Ξ32 (about $35,000
as of publication) to become a validator, Whereas Cardano allows people to delegate
their stake and participate with as little as 2.17 ADA.32,33

Figure 6
PoS Blockchain Validator Node Specs

Capped
PoS Validator Min. Stake Min. Stake Slashing
Bloackchain Set CPU Cores Ram (GB) Storage (Native) (USD) Penalty
Ethereum 2.0 - 4 pCPU 8 500 GB, Ξ 32 $38,207 ✓
SSD

Cosmos 125 4 pCPU 32 2 TB, 37,801 $340,133 ✓

SSD ATOM*

Polkadot 297 8 pCPU 64 500 GB, 1.85M $12,425,525 ✓

SSD DOT*

BSC 21 8 pCPU 16 1 TB, 614,846 $146,187,997 ✓

SSD BNB*

Solana - 12–16 pCPU 128–256 2 TB, 34 $1,238 ✓

SSD SOL

Algorand - 2–4 vCPU 4–8 100–200 0.1 $0.03

GB, ALGO
SSD

Avalanche - 8 pCPU 16 1 TB, 2,000 $39,420

SSD AVAX

Source: Kraken intelligence, protocol white papers, block explorers

August 2022 For more Kraken Intelligence content: kraken.com/subscribe/intelligence Page 23

 • The trade-off for eliminating the need for validators to purchase and continuously
upgrade hardware is that staked assets on PoS networks present an opportunity
cost requiring a participant to own the native asset to stake.
• Forking is not automatically discouraged by PoS systems as it is in PoW due to
the nothing-at-stake problem. However, PoS protocols have mitigated these
vulnerabilities by destroying the stake of dishonest validators and barring them
from the validation process. Still, readers should take such considerations into
account since some PoS networks, including Algorand and Avalanche, do not
include slashing penalties.34,35

The Debate: PoW vs. PoS

The rivalry between PoW and PoS surfaces key questions of network security,
sustainability, barriers to entry, and achieving decentralization. A series of trade-offs
between scalability, decentralization, and network security, commonly known as the
"blockchain trilemma," plague every blockchain. For example, an increase in throughput
typically comes at the expense of decentralization or security. Neither Sybil resistance
mechanism is perfect; it is essential to understand the way we clearly define these trade-
offs before discussing an optimal design for a blockchain network:

• Decentralization—retaining low barriers to entry for participation and a

sufficiently distributed voting system.
• Security—ability to conduct attacks against or manipulate the network.
• Scalability—transaction efficiency and ability to adapt to a growing, global audience.

Considering these definitions, it is our take that PoW tends to offer better security and
decentralization guarantees, sacrificing scalability in the process. On the other hand, PoS
typically offers better scalability in exchange for security and decentralization.

August 2022 For more Kraken Intelligence content: kraken.com/subscribe/intelligence Page 24

Figure 7
The Blockchain Trilemma–PoW vs. PoS

Criteria PoW PoS

Decentralization ✓

Security ✓

Scalability ✓

Source: Kraken Intelligence

However, the optimal choice ultimately depends on a given blockchain's use case and how
developers implement its Sybil resistance mechanism. Some blockchains are better suited for
PoS, while others should use PoW.

Blockchain networks should generally adhere to a PoW mechanism if they want to retain the
ethos of crypto: decentralization and security. PoW is generally more secure than PoS in that
it is more extensively vetted, has fewer potential attack vectors (e.g., bribery attacks), requires
the use of both capital and labor to misbehave on the network, and discourages constant
forking (i.e., nothing-at-stake). The mechanism is also more decentralized than PoS in that
it inherently encourages miners to distribute globally in search of cheap energy sources.
Furthermore, voting power is theoretically less susceptible to falling in the hands of the
wealthiest cryptoasset holders due to the requirement of hardware in authoring new blocks.

Imagine a scenario where a mining pool operator acts maliciously and the mining pool
contributors opt to leave the pool. At worst, the malicious mining pool operator could
keep the contributor’s earnings. However, they could not continue to utilize their hash
power for malicious purposes because the individual miners could turn off their mining
equipment. Conversely, suppose a similar situation emerges where hackers target a third-
party custodian and utilize the stolen funds to gain an overwhelming influence over a PoS
network. Unlike PoW, in this hypothetical situation, the customers of the compromised
custodian could not withdraw their contributions, losing both their funds and voting
power.

August 2022 For more Kraken Intelligence content: kraken.com/subscribe/intelligence Page 25

Thus, for use cases such as hard money, PoS likely is undesirable because the possibility
of the wealthiest network participants gaining an overwhelming share poses significant
problems for an asset whose value derives from its decentralization, security, and
scalability, among others.

PoS-based blockchains have more potential to scale because block production does not
consume energy, honest nodes can bar dishonest nodes from the validation process, and
the lack of mining allows for a faster validation process with less block variance. For use
cases including mediums of exchange or smart contract platforms, PoW is potentially less
desirable than PoS because network efficiency and scalability are paramount. Suppose
these blockchains were to prioritize decentralization and security. In that case, they might
become vulnerable to long-term scaling issues critical for a network that requires high
transaction throughput to scale for a global audience. Solana is a fine example of this.
Solana processes an average of 2,700 transactions per second (TPS), per the Solana explorer,
with an upper peak of over 710,000 TPS.36,37 Despite the immense transaction throughput
on this blockchain network, some argue that running a Solana node is infeasible for
consumers, leading to a centralized network. The high barrier to entry for participation
means anyone interested in running a Solana node needs data center-grade hardware,
including at least 128-256 GB of RAM, 2 TB of storage on an SSD, and 16 CPU cores.38
Furthermore, the high throughput has caused security concerns because the protocol has
broken down during times of high congestion.9

August 2022 For more Kraken Intelligence content: kraken.com/subscribe/intelligence Page 26

Figure 8
PoW (Top) vs. PoS (Bottom)—Transaction Count (7-Day Moving Average)

Daily Transaction Count (PoW)

PoW Asset Jun 20, 2021 Jan 1, 2022 Jun 13, 2022 YoY YTD

BTC 222.3K 241.7K 252.6K 14% 5%

LTC 93.3K 100.4K 97.3K 4% -3%

DOGE 23.8K 22.5K 23.1K -3% 3%

ETH 1.2M 1.2M 1.0M -10% -15%

BCH 92.8K 47.3K 41.2K -56% -13%

Total 1.6M 1.6M 1.5M -8% -11%

August 2022 For more Kraken Intelligence content: kraken.com/subscribe/intelligence Page 27

Daily Transaction Count (PoS)

PoS Asset Jun 20, 2021 Jan 1, 2022 Jun 13, 2022 YoY YTD

ICP 2.9K 32.2K 31.9K 984% -1%

ADA 27.5K 39.5K 83.7K 205% 112%

ALGO 3.4M 12.7M 5.1M 51% -59%

DOT 107.4K 179.7K 102.6K -4% -43%

OMG 0.7K 0.4K 0.2K -66% -48%

Total 3.5M 12.9M 5.4M 52% -59%

Source: Kraken Intelligence, Coin Metrics

August 2022 For more Kraken Intelligence content: kraken.com/subscribe/intelligence Page 28

Data suggests PoS cryptoassets are gaining traction as a medium of exchange as they
outperformed PoW cryptoassets in transaction count on an annual basis. However, PoS
cryptoassets have underperformed compared to PoW on a relative basis so far in 2022.
Demand for on-chain value transfer has grown, as evidenced by BTC’s transaction growth
on both a YoY and YTD period, and LTC’s YoY rise. ALGO’s poor performance primarily
drove the -60% YTD decline posted in the PoS camp. Because ETH alternatives like ADA and
ALGO saw network activity rise while ETH saw activity fall, this provides further evidence
of rising demand for smart contract platforms with low fees.

Furthermore, PoW cryptoassets have lost significant market dominance while token
and PoS dominance has grown over the last five years. PoS cryptoassets have consumed
considerable market dominance, but PoW still makes up roughly 58% of total market
dominance. Still, PoS cryptoassets are on track to eventually outgrow the market
capitalization of PoW cryptoassets should this pace continue. Moreover, readers should
note that once Ethereum undergoes the Merge, its transition from PoW to PoS, about
12% dominance will leave the PoW camp for the PoS camp, meaning PoW assets will
constitute about 46% of market dominance. In comparison, PoS will make up about 24%.

August 2022 For more Kraken Intelligence content: kraken.com/subscribe/intelligence Page 29

Figure 9
PoW vs. PoS Market Capitalization and Dominance

Source: Kraken Intelligence, CoinGecko, Project websites

Note: The readings in figure 6 track nearly 115 different cryptoassets (~0.85% of cryptoassets), making up more than 94% of the total crypto market
capitalization as of June 2022.

The choice between PoW and PoS is not black and white; however, there are apparent
differentiating factors that could help determine which Sybil resistance mechanism is
preferable for a given blockchain. Readers should interpret these comparisons as a rule of
thumb rather than an objective fact applicable across all blockchains. Every blockchain
still requires a deep understanding of the protocol’s measures to defend against Sybil
attacks, achieve a consensus on the state of the network, and how its measures balance
the trade-offs within the blockchain trilemma.

August 2022 For more Kraken Intelligence content: kraken.com/subscribe/intelligence Page 30

4.
Conclusion and outlook

A blockchain's consensus method, paramount for verifying the authenticity of distributed

blockchain platforms, is the process of building agreement among a network of mutually
distrusting participants. Maintaining BFT within an open and distributed network as
grand as Bitcoin requires a specific set of rules and mechanisms that rely on cryptography
and game theory mechanics to create the trustless environment necessary to facilitate
decentralized consensus across a value transfer network. The Sybil resistance mechanism
is arguably the most critical feature of a blockchain's consensus method since distributed
networks cannot otherwise reach a consensus in a BFT way. The two main Sybil resistance
mechanisms to maintain BFT in a large distributed system like Bitcoin or Cardano are PoW
and PoS, each with their trade-offs.

The rivalry between PoW and PoS confronts key questions of network security,
sustainability, barriers to entry, and decentralization. Understanding the trade-offs
between each is essential before concluding the optimal choice for a given blockchain
network. It is our take that PoW generally offers better security and decentralization
guarantees, exchanging scalability in the process. Conversely, PoS typically offers better
scalability in exchange for security and decentralization. However, the best choice
ultimately depends on a given blockchain’s use case.

Blockchains should adhere to a PoW mechanism if they want to retain the ethos of
cryptoassets: decentralization and security. For use cases such as hard money, PoS is likely
less desirable because the possibility of the wealthiest users gaining an overwhelming
share poses significant problems for an asset whose value derives from its decentralization,
security, and scalability, among others. Decentralization and security should take
precedence in these blockchains if they are to scale globally.

July 2022 For more Kraken Intelligence content: kraken.com/subscribe/intelligence Page 31

For use cases including mediums of exchange or smart contract platforms, PoW is
potentially less desirable than PoS because network efficiency and scalability are
paramount if these types of blockchains are to resist long-term scaling issues. Thus, the
choice between PoW and PoS is not black and white. It requires a nuanced understanding
of the two and their trade-offs to determine which Sybil resistance mechanism is better
suited for a particular blockchain.

August 2022 For more Kraken Intelligence content: kraken.com/subscribe/intelligence Page 32

Footnotes

1. https://en.wikipedia.org/wiki/Sybil_(Schreiber_book)

2. https://ethereum.org/nl/developers/docs/consensus-mechanisms/

3. https://dl.acm.org/doi/10.1145/357172.357176

4. https://cypherpunks-core.github.io/ethereumbook/14consensus.html

5. https://arxiv.org/abs/2203.01315

6. https://arxiv.org/abs/1710.09437

7. https://www.leewayhertz.com/solana-blockchain-using-poh/#:~:text=Solana%20runs%20a%20consensus%20
mechanism,out%20of%20voting%20on%20an

8. https://link.springer.com/chapter/10.1007/3-540-48071-4_10

9. https://link.springer.com/chapter/10.1007/978-0-387-35568-9_18

10. https://www.saveonenergy.com/electricity-rates/#:~:text=The%20average%20residential%20electricity%20
rate,kilowatt%2Dhour%20(kWh).&text=The%20average%20electric%20price%20a,is%2011.77%20cents%20per%20kWh.

11. https://bitcoinmagazine.com/technical/now-segwit2x-hard-fork-has-really-failed-activate

12. https://bitcoinmagazine.com/technical/no2x-hard-fork-suspended-due-lack-consensus

13. https://bitcoin.stackexchange.com/questions/273/how-does-merged-mining-work

14. https://en.wikipedia.org/wiki/Bitcoin_Gold

15. https://www.coindesk.com/markets/2017/11/08/ethereum-security-lead-hard-fork-required-to-release-frozen-parity-
funds/

16. https://coin.dance/blocks/hashrate

17. https://link.springer.com/chapter/10.1007/978-3-662-53357-4_2

18. https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=8653269

19. https://www.coindesk.com/markets/2020/08/29/ethereum-classic-hit-by-third-51-attack-in-a-month/

20. https://www.ccn.com/privacy-coin-verge-succumbs-to-51-attack-again/

21. https://blockexplorer.com/news/third-times-a-charm-verge-suffers-51-attack-yet-again/

22. http://fortune.com/2018/05/29/bitcoin-gold-hack/

23. https://www.ccn.com/vertcoin-hit-by-51-attack-allegedly-lost-100000-in-double-spending/

August 2022 For more Kraken Intelligence content: kraken.com/subscribe/intelligence Page 33

24. https://btc.com/stats/pool

25. https://www.cs.cornell.edu/~ie53/publications/btcProcFC.pdf

26. https://www.coindesk.com/tech/2021/03/11/nanos-network-flooded-with-spam-nodes-out-of-sync/

27. https://cointelegraph.com/news/solana-hit-with-another-network-incident-causing-degraded-performance

28. https://www.peercoin.net/#:~:text=Efficient%20Security&text=The%20key%20innovation%20of%20Peercoin,a%20
costly%20limited%20resource%3A%20electricity.

29. https://blog.ethereum.org/2021/05/18/country-power-no-more/

30. https://arxiv.org/pdf/2003.03052.pdf

31. https://decrypt.co/21108/did-binance-just-help-take-over-steem-network-justin-sun

32. https://ethereum.org/en/staking/

33. https://www.reddit.com/r/Cardano_ELI5/comments/lg70t3/comment/gn6o7rl/

34. https://twitter.com/el33th4xor/status/1304204689198723073?lang=en

35. https://medium.com/blockchain-at-berkeley/the-need-for-an-incentive-scheme-in-algorand-6fe9db45f2a7

36. https://explorer.solana.com/

37. https://docs.solana.com/introduction

38. https://docs.solana.com/running-validator/validator-reqs

39. https://www.coindesk.com/tech/2022/05/03/heres-why-solana-ceased-block-production-for-seven-hours-on-
saturday/#:~:text=Solana%20processes%20an%20average%20of,network%2C%20as%20per%20developer%20
documents.

August 2022 For more Kraken Intelligence content: kraken.com/subscribe/intelligence Page 34

 We appreciate your feedback!

Please visit here to participate in a brief survey. For all future Kraken Intelligence
content, sign up here. For comments, suggestions, or questions related to this article or
future topics you’d like to learn more about, you may also direct your communication
to [email protected] or to your account manager.

Kraken provides access to 196 crypto assets spanning 601 markets with advanced
trading features, industryleading security, and on-demand client service. With the
acquisition of Crypto Facilities, Kraken now offers seamless access to regulated
derivatives on 5 cryptocurrencies with up to 50x leverage. Sign up for a free account in
minutes at www.kraken.com/sign-up. We look forward to welcoming you.

For multi-exchange charting, trading, portfolio tracking, and high resolution historical
data, please visit https://cryptowat.ch. Create a free Cryptowatch account today at
https://cryptowat.ch/account/create.

For OTC-related execution services or inquiries, please direct your communication to

[email protected] or to your account manager.

Disclaimer
The information in this report is provided by, and is the sole opinion of, Kraken’s research desk. The information is
provided as general market commentary and should not be the basis for making investment decisions or be construed
as investment advice with respect to any digital asset or the issuers thereof. Trading digital assets involves significant
risk. Any person considering trading digital assets should seek independent advice on the suitability of any particular
digital asset. Kraken does not guarantee the accuracy or completeness of the information provided in this report,
does not control, endorse or adopt any third party content, and accepts no liability of any kind arising from the use of
any information contained in the report, including without limitation, any loss of profit. Kraken expressly disclaims all
warranties of accuracy, completeness, merchantability or fitness for a particular purpose with respect to the information
in this report. Kraken shall not be responsible for any risks associated with accessing third party websites, including
the use of hyperlinks. All market prices, data and other information are based upon selected public market data, reflect
prevailing conditions, and research’s views as of this date, all of which are subject to change without notice. This report
has not been prepared in accordance with the legal requirements designed to promote the independence of investment
research and is not subject to any prohibition on dealing ahead of the dissemination of investment research. Kraken and
its affiliates hold positions in digital assets and may now or in the future hold a position in the subject of this research.
This report is not directed or intended for distribution to, or use by, any person or entity who is a citizen or resident of, or
located in a jurisdiction where such distribution or use would be contrary to applicable law or that would subject Kraken
and/or its affiliates to any registration or licensing requirement. The digital assets described herein may or may not be
eligible for sale in all jurisdictions.

August 2022 For more Kraken Intelligence content: kraken.com/subscribe/intelligence Page 35