4 Misconceptions about PoS vs PoW

Dear Bankless Nation,

The merge is coming this summer.

In one fell swoop, the merge will:

• Reduce ETH’s issuance to net-0% and likely deflationary depending on EIP-1559 burn rate
• Increase network security
• Drop energy consumption by an estimated 99.95%
• Set the stage for an L2-centric scalability plan

Ethereum is critical infrastructure for the future of the web, financial markets, digital asset ownership, work, identity, and more.

It’s not every day such infrastructure sees a collective software upgrade.

However,

Moving from PoW to PoS is a massive feat that has technical, monetary, economic, and social ramifications. So of course there will a combination of miscommunication, FUD, and criticism.

Today, David sets the record straight on 4 major misconceptions about PoS and PoW.

- RSA

The Merge Cometh

Later this year, the Proof of Work Ethereum blockchain will ‘merge’ with the existing Proof of Stake Beacon Chain. This upcoming network upgrade will remove PoW from Ethereum and replace it with the PoS consensus mechanism that currently secures the Beacon chain.

This transition to Proof of Stake will be the most significant blockchain network upgrade this industry has seen, and probably ever will see.

As the Merge approaches, the conversation around Proof of Work vs Proof of Stake will only become more relevant, especially as crypto networks increasingly integrate themselves into everyday life.

Consensus mechanisms are complicated, and the process of understanding them is riddled with rabbit holes and mental traps. The crypto industry itself hasn’t agreed on what is true about the nature of consensus mechanisms, the lasting impacts on their respective ecosystems, and the assets which power them.

This article will clarify some of the frequent misconceptions about Ethereum’s Proof of Stake, and simultaneously illustrate some highly underappreciated strengths of a PoS consensus mechanism.

Table of Contents

Opener: PoW is PoS with Extra Steps

• People think PoW and PoS are super different. While they have different properties, they still ultimately collapse down to the same thing.

1. Which has higher returns on capital?

• Which system has higher returns on investment? Why does that matter?

2. What Type of Money does PoW vs PoS Produce?

• Does the consensus mechanism impact the nature of the asset?

3. Governance & Power differences in PoW vs PoS

• How does the consensus mechanism impact the governance structure of the blockchain? Does Proof of Stake give governance powers to stakers?

4. Dealing with 51% Attacks

• How does PoW vs PoS recover from 51% attacks?

Conclusion: Which one is “legacy tech”

• Hint: It’s not the one that took years of research

1. Which has higher returns on capital?

• Which system has higher returns on investment? Why does that matter?

PoS is often critiqued as the “rich get richer” system, because of its similarity to compounding interest on your financial assets. “If you’re rich, you earn capital faster than if you’re poor!”

But this is completely backward. By making ETH the capital asset that provides security, you’re able to control for this. Proof of Stake returns are programmatically guaranteed to provide the same rate of return for all stakers; having more ETH does not yield a higher rate of return.

In contrast, having more capital does yield a higher RoR in PoW.

If two Bitcoin mining facilities are built, one with a $100k of investment and the other with $100M, the $100M facility will produce significantly more than 100x the hashpower than the $100k facility. This is because of the extra steps that separate $1 of capital from becoming $1 worth of Bitcoin hashes. Purchasing hardware, hardware decay, electricity consumption, building the facility, cooling the facility… the list goes on. Every step that separates $1 of capital from $1 of hash rates is a vector upon which economies of scale can be expressed, giving the advantage to well-capitalized entities over the modest individual.

By eliminating these steps, PoS becomes the most globally equitable consensus mechanism for public blockchains.

It’s actually PoW that’s the rich-get-richer system. PoS enables you to stake ETH at home, but PoW forces you into a competition of economies of scale, which both centralizes the security and concentrates the capital into a new oligopoly of network-controlling players.

That’s legacy tech.

$1M investment into ETH staking will net the same amount of compound interest as a $1,000 investment, and the same as a $1B.

That’s Sci-Fi tech.

There are some efficiencies of scale in PoS. Larger validators can afford more frequent transactions to increase long-term yield rates. As stake increases, the costs of monitoring the validators also increases. Having a better computer & internet is probably a good expense.

But the delta between large and small stakers pales in comparison to the delta between large and small miners.

Proof of Work produces structural imbalances between elites and the masses, which is exactly the dynamic this industry is built to eliminate.

2. What Type of Money does PoW vs PoS Produce?

• Does the consensus mechanism impact the nature of the asset?

There’s a narrative that the formal association with energy consumption turns PoW currencies into “commodity-money”. Similarly, the “use the asset to get dividends on the asset” turns PoS currencies into “equity-money”.

I think these are interesting conceptual models, but they don’t actually produce any empirical definitions.

ETH has both commodity-like and equity-like characteristics.

BTC mostly has neither.

Commodity

The “commodity” that is produced by blockchains is their blockspace. Apple sells iPhones, Google sells ads. Blockchains sell blocks.

It costs money to buy transaction-space on a blockchain (gas), and you need to use the respective currency of each blockchain in order to purchase blockspace.

There is an association between the blockspace of a chain, and the respective currency that can purchase it. BTC is the only currency that can purchase Bitcoin blockspace, and ETH is the only currency that can purchase Ethereum blockspace. This places some amount of demand for the respective currencies as a function of the demand of each chain's blockspace.

This makes the nature of the blockspace a part of the nature of the asset. Since the asset has a monopoly on the blockspace, the characteristics of the blockspace become imbued into the nature of the asset.

Here’s a key difference between Bitcoin and Ethereum:

With Bitcoin, there is no formal relationship between the blockspace and the asset, other than that you must use BTC to purchase Bitcoin blockspace.

In contrast, Ethereum's EIP1559 directly associates the value of Ethereum’s blockspace to the value of ETH. EIP1559 transforms demand for Ethereum blockspace into demand for ETH the asset.

EIP1559 changes the relationship between the asset and the blockchain by consuming rather than recycling the currency that pays for blockspace. Purchasing Bitcoin blockspace sends BTC to the miners, which is then reintroduced on the secondary market (to cover mining costs). In contrast, purchasing Ethereum blockspace burns ETH (via EIP1559), consuming and removing it from the supply, like a commodity.

Ethereum blockspace supply is inelastic. The net effect of increased demand for blockspace results in an increased ETH consumption rate, since EIP1559 burns ETH at a rate that is a function of blockspace demand.

This imbues ETH with the commodity-value of Ethereum’s blockspace.

There generally isn’t much demand for Bitcoin blockspace, since the main utility of Bitcoin is holding BTC. Transfers need to be possible, but the value of Bitcoin does not lay in the transferring of BTC; it’s in the 21m hardcap supply.

Equity

PoWers make the comparison that PoS imbues ETH with equity characteristics. I do think that ETH has equity-like characteristics, but these properties do not come from the consensus mechanism. The price of ETH is formally linked to the value of the Ethereum network (establishing the ‘equity’ relationship) via two mechanisms:

• EIP1559
• Collateral demand in DeFi

Collateral

ETH is DeFi’s most pristine collateral; ETH’s trustless properties see more demand from DeFi than any other asset, making growth in DeFi result in an increase in ETH value. ‘Decentralized finance’ could also be called ‘trustless finance’, and in order to produce trustlessness in the early days of these expressions of finance, over-collateralization is required.

As DeFi applications grow in number and utility, more ETH is demanded as collateral. This induced reservation demand produces a relationship between the holistic DeFi superstructure and the value of ETH.

DeFi utility turns into ETH demand, linking the value of the Ethereum network to the price of ETH on the secondary markets. Ultimately, the price of Ether is a measure of the value of goods and services Ethereum provides, which sounds pretty equity-like.

Bitcoin doesn’t have the ability to host DeFi applications, and does not have this link between the network and the asset.

EIP1559

Similarly to imbuing ETH with commodity properties (as explained above), EIP1559 also imbues ETH with equity-like properties too.

All Ethereum applications produce generalized economic activity. Every application on Ethereum produces some incentive for its usage, creating demand for Ethereum blockspace. By consuming ETH for every transaction, EIP1559 links the value of economic activity to demand for ETH.

The impact that EIP1559 has on the nature of ETH the asset is both commodity-like and equity-like, and illustrates why these ‘equity-commodity’ comparisons are unhelpful.

Is ETH more a commodity or more equity? Both. Neither. It just is. Same with BTC. BTC isn’t a commodity. It’s 21 million units. It doesn’t share in the upside of the growth of the Bitcoin network, so it’s not equity either.

It just is what it is.

3. Governance & Power differences in PoW vs PoS

• How does the consensus mechanism impact the governance structure of the blockchain? Does Proof of Stake give governance powers to stakers?

Ethereum-flavored Proof of Stake is commonly conflated with explicit on-chain governance. There are blockchains with on-chain token governance like Tezos and Decred. Tezos is PoS, and Decred is hybrid PoW/PoS. There is no relationship between the consensus mechanism and token-governance; these are separate aspects of blockchain design that are frequently treated as the same.

Nothing about the ‘Work’ in PoW removes governance more than in PoS.

It’s not work that eliminates human governance from crypto-economic systems, it’s cryptography. Crypto uses cryptography to reduce human governance, not consensus mechanisms.

Lyn Alden once tweeted:

No, it's not 'work' to sign a Private Placement Memorandum for investment into a 150MW mining facility in order to mine BTC at a discount for the next decade, so that you can reinvest those profits into additional ASICs and additional mining facilities.

PoS does require work. Validators actually have to do real computations to move the network forward. Just like in PoW, the CPU of PoS validating computers heats up and puts out joules of energy as they process the transactions to embed in the blockchain. The difference is that all the computation of PoS validates is useful and not just a race to consume more electricity than others by producing meaningless hashes.

PoW doesn’t ‘ungovern’ itself by being resource-intensive. I’d argue the opposite.

Because of its resource intensivity, PoW subjects itself to ire from local governments and the spheres of influence around the hashpower. Because of PoW dependency on supply chains, physical buildings, energy consumption, and other physical things, it subjects itself to the current governance structures of the world that already govern over the physical world.

Crypto is here to build the future digital world; physical instantiations are a liability.

4. Dealing with 51% Attacks

• How does PoW vs PoS recover from 51% attacks?

PoW

With PoW, once someone amasses 51% of the hashpower, there’s no mechanism to prevent them from continuously and endlessly attacking the chain. This is what has been dubbed a Spawn-Camp Attack; an attacker can 51% attack the chain over and over again with the explicit goal of rendering it useless.

Under a 51% attack scenario, if honest miners cannot source additional hashpower then the attacker has full control over the transaction inclusion.

The hashpower that secures Bitcoin is produced externally from the network. ASIC’s aren’t ‘registered’ to Bitcoin in the same way that ETH is to Ethereum.

PoS

ETH is a virtual ASIC. Digital ASICs have some important advantages:

• ETH doesn’t decay; it strengthens. 

  • PoW-ers claim that this creates structural centralization, due to how validators on Day 1 are basically guaranteed to still be validators on Day 10,000, whereas miners are constantly churning due to the required investment in hardware, facilities, and operational costs. 

  • PoS-ers claim that this empowers the individual better, as the challenges of maintaining hashpower market share are only accessible to those with privileged knowledge, political connections, and large amounts of capital. PoS takes all these challenges, and contains them inside of the value of ETH, democratizing the upside of those benefits to those who are capable of simply holding and staking ETH. 

• ETH is infinity transportable, maximally discrete, and has no physical footprint

  • Staked ETH doesn’t actually exist anywhere in the real world. This makes the PoS network far more resilient to attack. PoW mining facilities are subject to physical forces (armies, governments, tanks) and cannot easily uproot themselves. The physical cumbersome nature of PoW facilities makes it susceptible to capture, just like how gold was ultimately captured by nation-states.

• Most importantly, ETH is known to Ethereum

  • Unlike ASICs, ETH is registered to the Ethereum network. The Ethereum network can identify between malicious vs honest stakers.

In a PoS attack, the attack is coming from a specific address that holds specific ETH that is being used to attack the network. The supply of staked ETH that would attack Ethereum is objectively viewable to everyone.

“The 15M staked-ETH at address 0xabc123attackeraddress789xyz proposed an invalid block, attempting to fork the network”

There are two possible scenarios here:

1. A staker produces an invalid block. The protocol can objectively view this invalid block, and slashes the staker who proposed it. No human intervention is required.
2. An attacker has >66% of the stake, and is able to prevent #1 from occurring. This allows the staker to censor all transactions.

In the dire circumstance of #2, PoS gives the system the capacity to fork into a new chain while deleting the stake of the attacker. The community identifies validators controlled by the attacker, and comes to a consensus about if and how to remove these validators from the active set, either by forcibly exiting them from staking, or even by removing these validators completely from the system, effectively burning the stake

Once this fork occurs, a significant amount of ETH is deleted from the network, making the proportion of ETH stake that is owned by honest, chain-aligned actors grow significantly. The ETH supply also goes down significantly, and everyone who didn’t attack the chain owns a larger share of the supply of ETH. It organically rewards those who were aligned with the network, while also increasing the costs for follow-up attacks, as the threshold for an attack is now a larger percentage of total supply. It’s really quite elegant.

This process isn’t a formal part of the Ethereum network (and may never even happen in practice, given the deterrence effects of the threat of a soft-fork). This requires human intervention and off-chain coordination to pull off; something that is generally against the ethos of cryptocurrency. The goal is to not have to fork. But, in contrast, Bitcoin doesn’t even have this option. In a PoW 51% attack, human intervention is also required to get the network to resume normal operation, but unlike in PoS, there’s no clear and obvious path towards getting honest hashpower back above 51%.

Both PoW and PoS result in external network actions in order to restore network consensus.

PoS:

1. Identify the address(es) of the attacker (trivial; just look on-chain).
2. Coordinate a block height to fork off the attackers stake (Non-trivial, but fork coordination has happened plenty of times in Ethereum’s history, and is a known quantity. We’ve done it before).

PoW

1. Coordinate with ASIC manufacturers and supply chains to ramp up ASIC production ASAP.
2. Once the ASICs are produced, ship the ASICs to mining facilities that have the resource capacity for them
3. Ensure that these new ASICs aren’t actually shipped to the miner doing the attacking (but you don’t know who it is!)

The timeline for fixing a PoW 51% spawn-camp attack is long. It’s impossible to say how long without knowing the nature of the attack and the resources of the attacker, but I think it’s fair to claim that it's easier to coordinate on a blockheight to fork a PoS network than it is to coordinate global supply chains to ship ASICs to specifically the hands of honest actors.

Not only is the PoW solution to a 51% attack a long process, but it also doesn’t produce any assurances that the solution will actually work. “Just produce more ASICs” is NOT a solution to a PoW 51% attack.

However, even more critically, #1 isn’t actually meaningfully incentivized by the network. Will miners actually take the risk of deploying capital to purchase ASICs while the chain is suffering a 51% attack? What if the BTC price is plummeting as a result of the loss of confidence in the chain? Will the miners actually receive the maximum ROI from the ASICs they need to purchase? Why take the risk? There are no guarantees that the network will ever come back online, seriously threatening the legitimacy of BTC the asset. If this causes BTC to fall in price, this removes even more incentive for honest miners to take the risk of investing more capital in producing honest hashpower.

Most importantly, this shows how control over a PoW chain comes down to who controls ASIC production and supply chains. Because of its relationship with electricity consumption, PoW chains have a commitment to how the atoms of the physical world are organized. Whoever controls the physical world controls the PoW chain.

This is the bargain that PoW makes; by being instantiated in the physical world, it cedes control to those who control the physical world.

Yes, Bitcoin mining facilities are spread out over the world, but this is in contrast to Proof of Stake validators which have no physical footprint. The ETH that is staked to the Ethereum network doesn’t actually exist on any one specific computer. The validating computer that is staking ETH can be destroyed, and the staked ETH can be recovered and redeployed from private keys anywhere in the world.

This is one of the many reasons why Proof of Stake is more decentralized than Proof of Work; no supply chain centralization, no economies of scale centralization, and no large mining facility centralization.

Conclusion: Which one is “legacy tech”

A common trope out of the Bitcoiner camp is that “Proof of Stake is legacy tech”.

The claim is that it's a reincarnation of the same structures of power the world already abides by, and that it’s actually Bitcoin & PoW that’s the real innovation.

I think anyone who makes that claim is seriously confounding critical elements of blockchain design. Conceptual models and metaphors about returns on capital or associations with legacy governance structure are no grounds for evidence, especially when they ignore the highly relevant details of Ethereum’s implementation of PoS.

One of my favorite critiques of Bitcoin culture is that when the base-chain doesn’t allow L1 innovation, it moves to the social layer. I think any claim that Proof of Stake, which has been rigorously researched by people on the frontier of crypto-economics, is “legacy technology” is simply spinning narrative and filling in the gaps of their knowledge with what they want to see.

Proof of Stake will go down in history as one of the most democratizing forces of power that has ever been invented.

This is the legacy that Ethereum is building for itself. The equitable distribution of money and power to the margins is best enabled by Proof of Stake, and the system has been designed to achieve these goals.

It’s the only way to maintain decentralization over the longest of time horizons.

- David