In a proof-of-work system, the hashpower that’s required is of a potentially unlimited supply. For instance: If a mining firm has 51% of current hashpower, their competitors can easily displace the mining firm’s dominance by acquiring ASICs and electricity. There is virtually no limit (within reason) to the amount of ASICs that can be produced or the amount of electricity that can be produced. Supporters of Bitcoin regard this quality of PoW as an asset and there is indeed some truth to it. But this means there is a theoretic unlimited amount of hash that can be weaponized against the protocol. It also fails to be a good solution when mixed with the centralizing effect of economies of scale (which Bitcoin miners benefit from, see part 4 for more detail).
In a proof-of-stake system, the main element of resilience to attacker accumulation is that acquiring stake is tied to the price of ETH. As an attacker buys the ETH necessary to mount such an attack, the slippage will become increasingly unsustainable. The fact that the resource (ETH) is scarce compared to the PoW resource (ASICs + energy) means that acquiring such a large stake in the network is far more difficult than acquiring majority hashrate because its a zero sum game. There is a bit of nuance here: PoW still has the centralizing effect of economies of scale (which make it hard to compete with majority hashrate), but large operations (of any variety of business) can acquire majority hash easier than they can acquire majority stake. The economies of scale aspect still holds true but so too does the relative ease of acquiring majority hashrate.
An ancillary reason supporting the previous paragraph is that, by nature, a lot of ETH will be locked in staking contracts. Validators won’t want to part with any of their ETH because they need it to earn a profit via validation. Currently there is ~13 million ETH staked out of ~122 million total ETH. But it’s certainly worth noting that there will most likely be much more ETH staked once the merge actually happens. And then even more once withdrawals are enabled. Some consider the risks associated with these two pending events to quite high and it discourages them as potential stakers. Liquid staking solutions are irrelevant to this equation because behind every liquid staking token is real ETH that is out of circulation and an attacker needs real ETH to attack the network.
In a PoW system an attacker can attack the network for little to no cost. There is no penalty for doing so and this can be attempted an indefinite amount of times in theory. Even if there were public outcry about a malicious miner executing many double spend attacks, the only resolution available is to change the hashing algorithm of Bitcoin, thereby rendering all mining equipment null (and fundamentally altering the ‘ossified’ protocol, which would require burnings and stakes but not of the Ethereum variety). Since hashpower can change IP addresses and point to different BTC deposit addresses, there is no way to surgically remove the malicious hashpower from the equation without altering the fundamentals of the entire protocol. Even after the hashing algorithm is changed, the same problem could eventually creep up again.
Due to the very low consequence of attacking BTC, renting out hashpower to attack becomes a very profitable business. The miner gets a guaranteed profit without any negative consequences.
PoS handles attacks in a much different way than PoW. In PoS since every validators’ ETH address is known, and since the validators need their stake to perform an attack, the solution to an objectively malicious validator is to slash their stake, thereby reducing their consensus power.
A recent collective brainstorm goes something like this: Coinbase holds 1/3rd of active stake. If they censor all sanctioned transactions via attesting against any blocks with sanctioned transactions, then social consensus becomes activated. There are two solutions, the first of which is an outright slash reducing their stake below the 1/3rd necessary to censor. The other solution is to create a softfork where their stake would be made inactive. This would consistently slash them for their inactivity until they are below the 1/3rd threshold, then kick them out of the validation set.
This is perhaps one of the less talked about benefits of PoS: that it’s harder if not impossible (with proper measures taken by validators) to physically locate where validation is occurring in order to shut it down.
Since PoW requires a lot of physical infrastructure and electricity to operate, it’s easier to locate and shut down if a governments decides to make mining or cryptocurrency illegal. Additionally governments can seize the physical infrastructure in order to commit their own mining operations. A recent example was when China banned Bitcoin mining and there was a mass exodus of miners from China. Where would they go if most countries banned it?
A PoS validator client runs normal software, so detection is extremely difficult if the validator is shielding their IP address via VPN routing. Even in the unlikely scenario that their physical location is compromised, their actual stake would be protected so long as they took protective measures regarding how they store their ETH. Sure, they would no longer be validating, but the government wouldn’t be able to seize the stake and use it for their own validation operation. The user could then run the PoS software on another computer, perhaps in another country, and continue staking as if nothing happened. But assuming that every validator can successfully hide their IP address, it would be nearly impossible to locate them and shut them down.
Since Bitcoin mining infrastructure is physical and in the real world (a feature not a bug according to Bitcoiners), it’s also capturable. Organizations/governments can physically seize all the mining operations. In the case that the miners are allowed to flee a certain jurisdiction: It’s quite difficult and expensive to move billions of dollars worth of mining infrastructure. However, it’s significantly easier to transport billions of dollars worth of stake.
While in theory anyone can amass PoW hashpower to competitively mine, it’s much more difficult to do so when large PoW miners benefit from economies of scale. Large mining operations get a higher percent yield on their investment than small mining operations. For instance: a $100 million investment in mining infrastructure will earn more percent return on investment than a $1 million investment. This is due to the nature of wholesale value, whether it’s in buying a lot of ASICs at one time, to electricity contracts, intellectual capital, etc. Having this benefit allows big mining operations to strangle and out compete smaller ones, trending toward centralization at the top.
In PoS everyone gets the same yield whether you’re staking 0.1 ETH (via liquid staking services) or 10 million ETH. However in PoW mining not only is the yield unequal depending on invested amount, but also earning a profit mining as an average person is extremely difficult. It’s hard to purchase the new ASICs when they are in such high demand and when they are shipping to the largest operations first. And if you don’t have new ASIC then your margins are destroyed by those who do.
Since miners requiring spending a large amount of money on electricity costs in order to sustain their operations, this makes the cost of security for Bitcoin higher than PoS. Miner profit margins are slim, and a vast amount of their total revenue goes toward electricity costs which are being subsidized via inflation. Even when the subsidization approaches 0, it will spell a security vulnerability for the network as it fails to accrue enough transaction fees to maintain a healthy security budget.
With PoS, since there is little to no operational cost (electricity), its essentially free for an ETH holder to become a validator. Therefore the protocol doesn’t have to subsidize validators with a lot of inflation, though some is still given for incentive to be a validator and secure the network.
Due to the extremely low cost of being a validator, we can assume that there will always be healthy security in PoS. Even when accounting for the opportunity cost of staking — assuming staking yield is only 1% for example, there are still liquid staking solutions allowing one to earn that 1% yield on stake and also utilize their liquid staking token. Whether that be selling it for another asset, lending it out, or using it as collateral for a more profitable venture.