Proof-of-stake (PoS) is a way for a blockchain to agree on which transactions are valid. Validators lock up the network's own coins as collateral instead of spending computing power. The network picks validators to propose and vote on blocks. Honest work earns rewards. Provable cheating destroys part of the stake. Ethereum, the largest proof-of-stake network, switched to this model on September 15, 2022. According to ethereum.org, the change cut its energy use by about 99.95%.
The short version
- Who secures the chain: validators who deposit stake. Ethereum requires 32 ETH to activate a validator.
- How blocks are chosen: one validator is picked pseudo-randomly to propose each block. Committees of other validators vote on it.
- Why cheating fails: provable misbehaviour, such as signing two conflicting blocks, is punished by "slashing" the cheater's deposit.
- How it differs from proof-of-work: proof-of-work spends electricity to make attacks expensive. Proof-of-stake puts the attacker's own capital at risk.
How proof-of-stake works on Ethereum
Becoming a validator
A validator deposits 32 ETH into Ethereum's deposit contract. It then runs three pieces of software: an execution client, a consensus client and a validator client. Since the Pectra upgrade went live on May 7, 2025 (epoch 364032), EIP-7251 lets a single validator hold an effective balance of 32 to 2,048 ETH. Before Pectra the cap was 32 ETH. Operators can now merge several validators into one.
Slots, epochs and votes
Time on Ethereum is divided into 12-second slots. Thirty-two slots make one epoch. In each slot the protocol uses a randomness process called RANDAO to pick one validator as block proposer. Every active validator casts an attestation once per epoch. An attestation is a signed vote for the block the validator considers correct.
Finality
The first block of each epoch is a checkpoint. A checkpoint becomes justified when validators holding at least two-thirds of all staked ETH vote for it. It becomes finalized when a later checkpoint is justified on top of it. This mechanism is called Casper-FFG. A second rule, LMD-GHOST, picks the chain with the most attestation weight when forks appear. Ethereum calls the combination of the two Gasper.
Once a block is finalized, reversing it would cost an attacker at least one-third of all staked ETH.
Rewards
ethereum.org says the base reward is proportional to the validator's effective balance. Each validator's share shrinks as more validators join. Rewards come from five duties, weighted out of 64:
| Duty | Weight |
|---|---|
| Timely target vote | 26 |
| Timely source vote | 14 |
| Timely head vote | 14 |
| Block proposal | 8 |
| Sync committee participation | 2 |
Missing a source or target vote costs about what the validator would have earned for it. ethereum.org states that there is no penalty for failing to propose a block.
Slashing and the inactivity leak
Two actions are slashable: proposing two blocks in the same slot, and submitting contradictory attestations. The penalty comes in three stages:
- An immediate penalty of 1/4096 of effective balance, up to 0.5 ETH.
- A "correlation penalty" around day 18. It can be below 0.1% for a lone validator and up to 100% if many validators are slashed together.
- Ejection from the validator set around day 36.
If the chain goes more than four epochs without finalizing, the inactivity leak starts. Validators that are offline or voting against the majority slowly lose stake until active validators again control more than two-thirds of the total, and finality can resume.
Which Ethereum staking route matches your situation?
Answer up to three questions to see which of the staking options listed by ethereum.org matches your ETH amount and how much you want to run yourself.
Do you have 32 ETH you are prepared to lock up as a validator deposit?
- Yes, 32 ETH or more → next question
- No, less than 32 ETH → next question
Do you want to run the three pieces of validator software yourself: an execution client, a consensus client and a validator client?
- Yes, I will run them myself → Solo (home) staking
- No, I want an operator to run them → Staking as a service
Which risk would you rather carry: smart-contract and counterparty risk, or trusting a centralised platform to hold your stake?
- Smart-contract and counterparty risk → Pooled or liquid staking
- Trusting a centralised platform → Custodial staking on a centralised platform
Needs 32 ETH. You keep full control of your keys, but you face penalties if your validator is offline and slashing if it is misconfigured.
Needs 32 ETH. An operator runs the validator for you and usually holds your signing keys.
Needs roughly 1.5–4 ETH per validator. You share operation of the validator with a protocol.
Works with any amount, some from about 0.01 ETH. The trade-off is counterparty and smart-contract risk.
Works with any amount. ethereum.org says this option carries the "highest trust assumptions".
This only maps your answers to the options ethereum.org lists for Ethereum; it does not cover rewards or yields, tax, other proof-of-stake chains, or whether staking suits you at all.
Where proof-of-stake came from
The first live proof-of-stake currency is usually credited to Peercoin. Its paper, by Sunny King and Scott Nadal and dated 19 August 2012, describes a hybrid design. Stake, weighted by "coin age," provides most of the network's security. Proof-of-work is used mainly for the initial creation of coins.
Early chain-based designs like this had a weakness known as nothing-at-stake. Vitalik Buterin's Proof of Stake FAQ (2017) explains that they offered "only rewards for producing blocks, and no penalties." That made it rational for validators to build on every competing fork at once. The fix was slashing: penalising validators who sign blocks on multiple chains. Ethereum's design builds directly on that fix.
Benefits and trade-offs
What ethereum.org lists as advantages
- Energy: about 99.95% less energy used after the Merge.
- Lower hardware barrier: no specialised mining equipment. Pooled options let people take part with less than 32 ETH.
- Crypto-economic security: attacks put the attacker's own stake at risk of destruction.
- Lower issuance: less new ETH is needed to pay for security.
What ethereum.org lists as drawbacks
- Proof-of-stake is younger and less battle-tested than proof-of-work.
- It is more complex to implement.
- Running a validator means operating three pieces of software.
Security thresholds
ethereum.org describes what an attacker could do with different shares of the total stake:
- 33%: stop the chain from finalizing.
- 34%: cause two conflicting blocks to be finalized, but only by committing slashable offences that destroy the attacker's stake.
- More than 50%: dominate fork choice, censor transactions and reorganise recent blocks.
- 66%: finalize a chain of the attacker's choosing.
If an attack succeeds, ethereum.org says the "ultimate fallback" is the social layer, meaning the community coordinating off-chain.
Weak subjectivity
ethereum.org states that "Subjectivity is inherent to proof-of-stake blockchains." A new node, or one that has been offline a long time, needs a trusted recent state, called a weak-subjectivity checkpoint, to join safely. Without it, the node could be misled by an old fork that early validators kept and released much later (a long-range attack). The documentation calls this trust requirement low-risk because the checkpoint can be compared across several nodes or block explorers.
Concentration of stake
Large holders and large staking pools get proportionally more influence over block production. ethereum.org notes that concentrating stake in big pools creates "a large centralized target and point of failure."
Ways to take part in staking on Ethereum
ethereum.org lists these options, ordered from most to least control:
| Option | ETH needed | Main trade-off |
|---|---|---|
| Solo (home) staking | 32 ETH | Full control of keys; penalties if offline, slashing if misconfigured |
| Bonded node operation | Roughly 1.5–4 ETH per validator | Shared operation with a protocol |
| Staking as a service | 32 ETH | An operator usually holds your signing keys |
| Pooled or liquid staking | Any amount (some from about 0.01 ETH) | Counterparty and smart-contract risk |
| Custodial staking on a centralised platform | Any amount | ethereum.org calls this the "highest trust assumptions" |
Limitations: protocol details come mainly from Ethereum's own documentation, and other proof-of-stake chains use different rules, parameters and penalties.
