Staking is the act of locking up crypto as collateral to participate in a proof-of-stake network’s consensus mechanism. If you want to be a validator on Ethereum, you deposit 32 ETH into the official staking contract and run software that proposes and attests to blocks. The 32 ETH is your bond — it is at risk of being partially or fully slashed if you misbehave, and it earns you staking rewards (currently around 3-4 percent per year on Ethereum) for participating honestly. The rewards come from a combination of newly-issued ETH and transaction fees, and they represent the network’s payment to you for securing it.
This is the narrow, technical meaning. The word “staking” has been stretched in marketing to cover a much broader range of lock-up-for-yield arrangements on protocols that do not actually use the stake for consensus — more on that in a moment — but the core concept is this: you lock up tokens, you participate in securing the network, you get paid for it, and you face slashing if you misbehave.
How Ethereum Staking Actually Works
Ethereum’s staking system launched in December 2020 with the Beacon Chain, initially as a separate chain that ran in parallel to the proof-of-work mainnet. When the Merge happened in September 2022, the Beacon Chain’s validators took over block production on mainnet, and Ethereum became fully proof-of-stake. As of 2026, there are around a million active validators securing the network, representing roughly 35 million ETH staked — close to 30 percent of the total supply.
Running a validator requires: 32 ETH, a server with reasonable hardware (the minimum is modest but 24/7 uptime matters), a consensus client (Lighthouse, Prysm, Teku, Nimbus, or Lodestar), and an execution client (Geth, Nethermind, Besu, Erigon, or Reth). You run both clients, they talk to each other, and you follow the network’s protocol correctly. For a competent person who cares about the technical setup, this is achievable at home on reasonable hardware. For everyone else, the barriers — technical complexity, the 32 ETH minimum, the operational discipline required to not get slashed — are high enough that most stakers participate through intermediaries instead.
Validator exits are partially rate-limited to prevent too much stake from leaving the network at once, which is a deliberate design choice to prevent mass exits from destabilising the chain. The exit queue can be days or weeks long depending on how many validators are trying to exit simultaneously, and this is a meaningful liquidity consideration if you are thinking about staking large amounts.
Liquid Staking and Staking Pools
Because running a validator is operationally demanding, most stakers instead use some form of pooled staking. The largest options:
Lido is a decentralised staking protocol that accepts ETH deposits and issues stETH (staked ETH) in return. The deposited ETH is distributed across a set of professional node operators who run the actual validators, and the rewards accrue to the stETH holders. stETH is freely tradeable on DeFi protocols, so you can stake and still use your capital for other purposes — this is the “liquid staking” model, and it is why Lido has become enormously dominant. Lido accounts for a significant percentage of all staked ETH, which has generated real concerns about centralisation.
Rocket Pool is a more decentralised alternative that lets individuals run their own validators with a smaller ETH requirement (8 or 16 ETH, supplemented by pooled ETH from other users) and issues rETH as the liquid staking token. Rocket Pool has maintained a more decentralised node operator set and is often recommended by people worried about Lido’s market share.
Coinbase, Binance, Kraken all offer staking services where they run the validators and pay rewards (minus a fee) to the depositors. These are simpler for users — you just click “stake” in the app — but you are giving up custody in exchange for convenience, and they have been the subject of SEC enforcement actions (Kraken had its US staking program shut down in 2023 as an alleged unregistered securities offering).
Solo staking is the ideological gold standard — you run your own validator, you keep all the rewards, and you contribute directly to network decentralisation — but it requires the 32 ETH minimum and actual technical work. The Ethereum Foundation and various community initiatives have pushed hard on making solo staking more accessible, but the market share trend has been the opposite, with more stake concentrating in pools and services over time.
The Broader “Staking” Misnomer
Outside of actual PoS consensus participation, the word “staking” gets used for all sorts of things it does not really mean. Many DeFi protocols call their “lock up this governance token for a while in exchange for rewards” mechanisms “staking” even though the locked tokens are not securing any consensus — they are just sitting in a contract and accruing emissions. Some exchanges call their yield products “staking” when they are really a form of lending. Some memecoin projects advertise “staking APY” for tokens that are nothing more than the team minting new supply and paying it out as “rewards” to holders who lock up.
These uses of “staking” are linguistically legitimate in the sense that the word has been colonised, but they do not have the same economic content as real PoS staking. Real staking produces a yield because you are performing a useful service (securing the network), and the yield is funded by block rewards and fees that represent real economic activity. Fake staking just recirculates token inflation back to the people who are holding the longest, and the yield is paid for by dilution of everyone else — a musical chairs arrangement where the exit door is small and the person who moves first wins. Knowing which kind of “staking” you are being offered is one of the more useful filters in DeFi, and the distinction is always worth looking into before locking capital.