Proof of Work vs Proof of Stake

Every blockchain has to solve one problem before it can do anything else: with nobody in charge, how do thousands of strangers agree on a single version of who owns what? Both proof of work and proof of stake answer it the same way in principle — make honest behaviour cheaper than dishonest behaviour — and completely differently in practice. One spends electricity. The other puts money at risk.

This page compares them on the things that actually differ. It is not an argument for either; the honest summary is that they buy similar security with different currencies, and which trade you prefer depends on what you think the currency is worth.

The short version

How proof of work secures a chain

To add a block, a miner repeatedly hashes its contents with a changing number until the result falls below a target. There is no shortcut — it is brute force, billions of attempts per second — but checking a winning answer takes an instant. That asymmetry is the entire mechanism: effort is expensive to produce and free to verify.

Because each block is expensive, so is rewriting one. An attacker wanting to erase a past transaction has to rebuild every block after it and out-run the honest network still extending the real chain. That is the “51% attack”: it needs a majority of the world's mining power, costs a fortune to sustain, and gets more hopeless the deeper the transaction is buried. Our Proof of Work page goes through this step by step.

The security budget is therefore denominated in energy. That is the criticism, and it is also the design: the cost is the protection, and removing it removes what it was buying.

How proof of stake secures a chain

Instead of spending electricity, validators lock up the network's own token as a deposit. The protocol selects who proposes the next block in proportion to how much is staked, and other validators attest that the block is valid. Behave correctly and you earn newly issued tokens and fees. Behave badly — sign two conflicting blocks, or go offline at scale — and the protocol destroys part of your deposit. That penalty is called slashing, and it is the mechanism doing the work that electricity does under proof of work.

Attacking the chain therefore means acquiring a majority of the staked tokens. There is an elegance to this that its supporters point out often: the attack requires owning a great deal of the very asset the attack would devalue, and the protocol can destroy your stake for trying. The cost of misbehaving is charged directly rather than paid to an electricity supplier in advance.

Running a validator on Ethereum requires 32 ETH. Smaller holders usually delegate to a pool or an exchange, which is convenient and quietly reintroduces the thing the design was supposed to avoid — concentration.

The criticisms, stated fairly

Against proof of work: it consumes a genuinely large amount of electricity to produce a number with no use outside the system, and mining has concentrated into large operations near cheap power, which is not the distributed picture the early literature described. Neither point is disputed by anyone serious; the argument is over whether the expense is worth what it buys.

Against proof of stake: influence is proportional to holdings, so the wealthiest participants have the most say and earn the most rewards — a structure critics argue drifts toward the concentration it was meant to escape. It is also younger under real adversarial pressure, and it depends on a more complex protocol, which means more surface for something to go wrong. Where proof of work anchors its security in a cost outside the system, proof of stake anchors it in the value of the token it secures, which is a circularity its critics find unconvincing and its defenders consider well-handled.

Why Ethereum switched and Bitcoin didn't

Ethereum moved from proof of work to proof of stake in September 2022, in an upgrade known as the Merge. It cut the network's energy consumption dramatically and was, by any measure, a substantial engineering achievement — a live network changing its consensus mechanism without halting.

Bitcoin has deliberately not followed, and the reason is cultural as much as technical. Bitcoin's value proposition rests on being hard to change: a network that alters its monetary mechanism because a better one came along is a network that could alter it again. The difficulty of changing Bitcoin is not friction its community tolerates — it is the feature they are paying for. Ethereum has always positioned itself as a platform that evolves, so an upgrade of this size fits what it claims to be. Neither project is being inconsistent; they want different things.

So which is better?

There is no answer that survives contact with the details, and anyone offering one confidently is arguing for a holding rather than explaining a mechanism. Proof of work buys security with a cost that is external, physical and verifiable, and pays for it in energy. Proof of stake buys security with capital at risk inside the system, and pays for it in complexity and in influence tracking wealth.

Both have now secured very large amounts of value for years. That is the most useful fact available, and it is the one both camps tend to skip past on the way to the argument they wanted to have.