Proof of Work vs Proof of Stake

Both answer the same question — who gets to add the next block, and why should anyone believe them — with completely different economics.

How proof of work secures a chain

Miners compete to find a number (a nonce) that, combined with the block's contents, produces a hash below a target. There is no method other than guessing, billions of times per second. The first to succeed broadcasts the block and collects the reward.

The security argument is economic. Rewriting history requires redoing all the work from the point of attack while simultaneously outpacing the honest network's ongoing production. That demands more than half the network's total hash rate, sustained — an expense in hardware and electricity that, for a large network, exceeds any plausible gain. An attacker who did acquire that capacity would also destroy the value of the asset they were attacking.

The cost is external to the system: real electricity, real hardware, real industrial operations. That is the feature and the criticism simultaneously.

How proof of stake secures a chain

Validators lock capital as collateral — 32 ETH on Ethereum — and are selected pseudo-randomly to propose and attest to blocks. Rewards accrue for honest participation.

The security argument is also economic, with the cost internalised. A validator who signs conflicting blocks or otherwise breaks the rules is slashed: part or all of their stake is destroyed by the protocol. Attacking the network requires acquiring a large share of the staked supply, and a successful attack results in that stake being destroyed.

The efficiency gain is dramatic. Ethereum's energy consumption fell by roughly 99.95 % when it switched in September 2022 — from the consumption of a mid-sized country to something comparable to a large office building.

An honest comparison

Both concentrate in practice. Proof of work concentrates into mining pools, where a handful coordinate most of the hash rate, and into regions with subsidised electricity. Proof of stake concentrates into large staking services and exchanges, where a few entities control a substantial share of validators. Neither community has solved this, and both tend to point at the other's version of it.

What is actually contested

Several disagreements are genuine and unresolved, and it is worth knowing which ones are real arguments rather than tribal noise:

• Does staking favour the wealthy structurally? Rewards accrue proportionally to stake, which critics argue compounds concentration. Defenders note that mining rewards accrue proportionally to capital too.
• Is externally-costly security stronger? Proof-of-work advocates argue that burning a real-world resource creates a harder-to-fake guarantee. Proof-of-stake advocates argue that destroyable capital is a more direct and better-targeted penalty.
• Nothing-at-stake and long-range attacks. Early theoretical objections to proof of stake; modern implementations address them with slashing and weak subjectivity checkpoints, though the latter introduces a small social trust assumption.
• Is the energy use justified? Depends entirely on how much you value the service. It is a value judgement wearing a technical costume, and arguments about the grid mix and use of stranded energy are real but do not settle it.

There is no consensus among researchers that one design is simply better. They make different trades, and the choice depends on what properties a given network is prioritising.