Proof-of-Stake and Staking

Bitcoin keeps itself honest by burning electricity: miners run warehouses full of chips, racing to win the right to add the next block, and the sheer cost of that race is what makes cheating pointless. Ethereum used to do the same. Then, in 2022, it pulled off one of the boldest engine swaps in crypto history — it stopped burning electricity to stay honest, and started burning something else: capital at risk. Instead of “prove you spent power,” Ethereum now says “prove you have money on the line, and watch what happens if you cheat.”

This lesson is about that swap — how proof-of-stake works, who validators are, why 32 ETH keeps coming up, what it means for a block to be final, and how ordinary people actually put their ETH to work securing the chain.

From Proof-of-Work to Proof-of-Stake

Both systems answer the same hard question: how does a leaderless network, with no boss and no trusted referee, agree on who gets to add the next block? They just answer it differently.

Proof-of-work — Bitcoin’s method, and Ethereum’s old one — makes participants compete with raw computing power. Miners run specialized chips that guess trillions of numbers per second, racing to solve a meaningless puzzle. The first to solve it wins the right to add the next block and collect the reward. The “proof” is the mountain of electricity you provably spent; faking it would cost as much as doing it honestly, so honesty is the cheap option.

Proof-of-stake flips the cost from energy spent to money at risk. Instead of burning power to earn the right to add a block, you lock up your own ETH as a bond (a security deposit). The protocol then picks participants — called validators — to do the work of building and checking blocks, and it picks them weighted by stake: the more ETH you’ve put on the line, the more often you’re chosen. Behave honestly and you earn rewards. Try to cheat in a way the protocol can prove, and it slashes you — destroys a chunk of your deposit and kicks you out. We’ll define slashing properly below; for now, hold the contrast.

Here’s the cleanest analogy. Proof-of-work is an arms race of power plants: the way you “prove” you deserve to add a block is by visibly torching resources, and security comes from the fact that attacking would cost a fortune in electricity. Proof-of-stake is a bonded contractor: before you’re allowed to touch the work, you post a security deposit, and if you do shoddy or fraudulent work, the deposit is forfeit. Same goal — make cheating more expensive than honesty — but one burns coal and the other puts your own savings on the chopping block.

The switch happened at a moment called The Merge (September 2022): Ethereum’s existing chain merged with a new proof-of-stake engine and abandoned mining overnight. The headline number: it cut Ethereum’s energy use by roughly 99.9% — from the consumption of a small country to that of a modest office building. Same ledger, same coins, same balances; an entirely new way of agreeing on them.

Validators and the 32-ETH Stake

A validator is the proof-of-stake equivalent of a miner: a participant the network trusts to help build and verify blocks. But where becoming a miner means buying hardware, becoming a validator means putting up money. To run one, you stake exactly 32 ETH. That deposit is your skin in the game — it’s locked into the protocol, it’s what you earn rewards on, and it’s what gets cut if you cheat.

Why a fixed 32 ETH per validator instead of “stake whatever you’ve got”? It keeps every validator the same size, which makes the math of selecting and rewarding them clean. If you control more ETH than that, you don’t make one giant validator — you run more validators, each with its own 32-ETH deposit. So more total stake controlled means more validators, which means more chances to be chosen to do work. Stake is influence, and it’s strictly proportional.

Play with the wheel below to feel it. Each colored slice is a validator, sized by its stake. Press Select a proposer several times in a row and watch which validators come up — over many presses, the share of times each one is picked tracks its share of the total stake (the biggest slice wins most often, the smallest rarely). Then press Slash a cheater: the chosen validator’s stake is cut, its slice shrinks, and everyone else’s chance rises — because chances are always relative to the total still on the line.

Proposing and Attesting

Proof-of-stake runs on a steady heartbeat. Time is sliced into slots of about 12 seconds each, and 32 slots make one epoch (roughly 6.4 minutes). Don’t over-memorize the numbers — just hold the rhythm: a slot is “one block’s turn,” and an epoch is the bigger bookkeeping cycle the protocol uses to tally votes and finalize history.

In each slot, two jobs happen:

• Propose. Exactly one validator is chosen (weighted by stake, as you just saw) to propose the slot’s block — to gather pending transactions, assemble them into a block, and broadcast it to the network.
• Attest. A committee of other validators is assigned to attest to that block — to check it and cast a vote that it’s valid and belongs on the chain. Attestations are the raw material of consensus: the chain everyone keeps building on is the one with the most honest votes behind it.

Do your job honestly and on time — propose a good block, or attest correctly — and you earn rewards. The work is light and continuous, not a winner-take-all race; almost everyone who behaves gets paid a little, all the time.

The analogy that sticks: think of a meeting. Each slot, one person is picked to draft the minutes (the proposer), and everyone else reads them and signs off that they’re accurate (the attesters). No single scribe can sneak a lie into the record, because the whole room has to vote it true — and the room is enormous and chosen unpredictably.

Rewards, Penalties, and Slashing

Proof-of-stake shapes behavior with a carrot and two very different sticks.

The carrot — rewards. Propose valid blocks, attest correctly, and stay online, and you earn a steady stream of ETH on top of your stake. It’s a salary for showing up and doing honest work, paid out continuously rather than as a jackpot.

The small stick — inactivity penalties. If your validator goes offline — server crash, internet outage — you stop earning and you bleed a little stake for each duty you miss. These penalties are gentle and symmetric to the rewards: roughly, you lose about what you’d have earned by being online. Annoying, recoverable, no catastrophe. The message is “please stay online,” not “you’re a criminal.”

The big stick — slashing. Slashing is the nuclear penalty, and it’s reserved for provable attacks on consensus. The classic example is equivocation (also called double-signing): signing two conflicting messages for the same slot — for instance, proposing two different blocks at once, or attesting to two contradictory versions of history. Because a signature is cryptographic proof of who did it, the network can catch the offender red-handed with no judgment call required. The punishment is brutal by design: a large chunk of the stake is destroyed and the validator is forcibly ejected. Worse, if many validators get slashed around the same time (a sign of a coordinated attack), the penalty scales up — the more it looks like an attack, the more it costs each attacker.

This is the whole security model in one sentence: attacking the chain means burning your own money. Under proof-of-work, an attacker rents hashpower and walks away with the hardware. Under proof-of-stake, the only way to attack is from inside, with staked ETH that the protocol can see and destroy — so a serious attack forces you to set fire to millions of dollars of your own capital, with nothing to show for it. Honesty isn’t enforced by morality; it’s enforced by arithmetic.

Finality

Finality is the question of how sure you can be that a block is permanent. Proof-of-work only ever gives you probabilistic finality: a block is never strictly final, it just gets exponentially harder to reverse as more blocks pile on top. That’s why Bitcoin users “wait for confirmations” — the deeper a transaction is buried, the safer it is, but you’re always playing the odds.

Proof-of-stake adds something stronger: economic finality. Validators don’t just build on blocks — they vote on them with their stake, and those votes get locked in at epoch boundaries. Once two-thirds of all staked ETH has attested to a span of history across two consecutive epochs, that history becomes finalized. From that point, reversing it isn’t merely improbable — it’s only possible if attackers are willing to get at least one-third of all staked ETH slashed. Reversing finalized history would, by construction, require so much provable equivocation that the protocol destroys a third of the entire stake in the process. So finality isn’t “very unlikely to be undone”; it’s “undoing it has a fixed, astronomical price tag attached, payable by the attacker.”

In short: proof-of-work says “the longer you wait, the safer you are.” Proof-of-stake says “after finality, undoing this costs an attacker a third of all staked ETH — so treat it as done.”

How People Actually Stake

Not everyone has 32 ETH and a server in the closet, so a few paths emerged — each trading control for convenience.

• Solo staking. You stake your own 32 ETH and run your own validator node — the purest, most decentralized option, with no middleman and the full reward. The cost: you need the 32 ETH, the technical chops, and a machine that stays online, or you eat inactivity penalties.
• Staking-as-a-service. You still put up your own 32 ETH, but you pay a provider to run the node hardware for you. Less hassle than solo, but you’re trusting the operator to behave and stay online.
• Pooled / liquid staking. You deposit any amount of ETH into a shared pool that runs many validators on everyone’s behalf, and in return you receive a liquid staking token — a tradable claim representing your staked ETH plus its accruing rewards. You stay liquid (you can sell or use that token in other apps while your ETH keeps earning) without ever locking up a full 32 ETH yourself.

The trade-offs sharpen as you move down the list. Anything but solo staking means trusting a third party with your validating, and pooled staking adds a centralization worry: if one liquid-staking provider attracts a huge share of all stake, it concentrates influence over consensus in one operator — the opposite of the decentralization the whole system is built on. And the liquid-staking token carries its own risk: it’s only as good as the operator behind it and the market’s faith in its peg, so it can trade below the value of the ETH it represents if confidence wobbles. Convenience is real, but it’s never free.

The Big Picture

Proof-of-stake in one frame: replace hashpower with capital at risk, pick validators by stake to propose and attest, pay them for honesty, slash them for cheating, and let stake-weighted votes finalize history. Chunk it:

A mixed recap pulling from every section:

Key Takeaways

You now know how Ethereum agrees on its ledger without burning a power plant’s worth of electricity. But proof-of-stake secures one chain at ~12 seconds a block — and that chain can only do so much work at once. Next we tackle how Ethereum scales beyond the base layer with Layer-2 rollups.