What Is Liquid Staking? How stETH Works, What It Pays and What It Costs

Staking ether has always involved a trade. You lock capital to earn a yield, and while it is locked you cannot spend it, lend it or sell it. Liquid staking exists to break that trade: you deposit ETH, receive a token that represents the same stake plus its rewards, and that token keeps trading while the underlying ether keeps validating. It is one of the few genuinely useful primitives in decentralised finance, and also one of the most misunderstood.

The misunderstanding is not about the mechanics, which are simple enough. It is about what the token actually guarantees. A liquid staking token gives you liquidity and a claim on a pool of staked ether. It does not give you a fixed redemption price, an instant exit, or immunity from the risks of the validators holding your coins.

The problem liquid staking solves

Ethereum's proof-of-stake design asks a validator to lock 32 ETH as collateral. Run your own validator and that 32 ETH sits in a deposit contract, earning rewards, unavailable for anything else. The capital is productive but frozen. The same is true at scale: as of early October 2026, roughly 43.7 million ETH was staked on Ethereum, about 35.8% of supply, spread across some 868,000 active validators.

For an individual with 4 ETH, solo staking is not even an option. For a fund with 40,000 ETH, solo staking means running 1,250 validators, or paying someone to do it. Liquid staking pools deposits of any size, hands the validation work to professional operators, and issues a transferable token in return. The ether is still staked; the receipt is liquid. Our guide to how staking works covers the underlying validator economics in more detail.

What liquid staking actually is

You deposit ETH into a protocol contract. The protocol allocates that ETH to validators it controls, and mints you a token representing your share of the pooled stake. From that moment, rewards accrue to the pool, and your token's value tracks your share of it.

Lido is the largest example and a useful reference point because its design is well documented. It launched at the end of 2020, holds roughly 9 million ETH staked, and accounted for 21.18% of all staked ether at the end of June 2026, according to its own half-year report. Its total value locked sat around $23 billion in late October 2026, the largest of any DeFi protocol, though well below the $33 billion it held a year earlier as ether's price fell.

The token you receive comes in two forms. stETH is a rebasing token: your balance increases every day as rewards are reported, so one stETH stays roughly equal to one ETH. wstETH is the wrapped version: your balance stays fixed and the token's exchange rate against ETH rises instead. The second form is what travels to layer-2 networks and sits in most DeFi contracts, because a balance that changes daily is awkward to account for in a smart contract.

Where the yield comes from

A staking reward is not one payment. Validators earn consensus rewards for attestations and block proposals, plus priority fees from the transactions they include, plus whatever extractable value a block builder can capture. All of it flows into the pool.

Lido takes a 10% fee on staking rewards, split between the node operators doing the work and the DAO treasury, with the exact split set per staking module. The remaining 90% accrues to stETH holders. That arithmetic matters more now than it did two years ago, because base staking yields have compressed as more ether has been staked. Bitwise's October 2026 staking report put Ethereum's annualised gross staking yield at 2.63%, a record low, which implies roughly 2.4% for a stETH holder after the protocol fee.

That is the honest headline: liquid staking does not create yield. It makes an existing yield liquid, and takes a cut for the service.

The peg is a market price, not a promise

stETH is designed to be redeemable for ETH one-to-one, but the token you hold trades on exchanges and in liquidity pools, where its price is set by buyers and sellers. When demand for an exit exceeds available liquidity, that price falls below the redemption value. This is not a hypothetical.

In May and June 2022, after the collapse of Terra's UST, stETH slid to a persistent discount. Three Arrows Capital pulled about $400 million of liquidity from the main stETH/ETH pool on Curve on 12 May 2022, Celsius froze withdrawals in June, and the Curve pool's stETH/ETH rate fell to roughly 0.94 at its low — a discount of about 6%. Holders who needed out immediately sold at that price. Holders who could wait eventually redeemed closer to par, because the underlying stake was never insolvent; the market for it was simply crowded.

That distinction is the whole lesson. A discount reflects a shortage of buyers or a shortage of liquidity, not necessarily a shortage of ether. It can also persist: stETH has traded at small discounts for long stretches since.

How you get your ETH back

For most of stETH's early life, you could not redeem it at all. The only exit was selling the token. That changed after Ethereum's Shapella upgrade on 12 April 2023 enabled validator withdrawals, and Lido's V2 upgrade went live on 15 May 2023 with in-protocol withdrawals.

The process today is a request-and-claim cycle. You submit a withdrawal request, which burns or locks your stETH and mints a transferable NFT representing the claim. The protocol sources ETH from its buffer or by exiting validators, then marks the request claimable, and you withdraw. Under normal conditions Lido quotes one to five days, and it can be faster when the buffer is full. It is not instant, and it is not independent of the network: requests queue behind Ethereum's own validator exit queue, which stood at roughly 786,000 ETH, or about 14 days, in early October 2026.

During the wait you stop earning rewards, and the claim is denominated in ETH rather than stETH. Anyone telling you liquid staking gives you instant liquidity is describing the secondary market, not the protocol.

Who actually runs the validators

This is the part that determines most of your real risk, and it is usually invisible on a dashboard. Lido runs a curated set of professional node operators, chosen and reviewed by the DAO, with performance and slashing records published. In October 2024 it added the Community Staking Module, a permissionless route for smaller operators who post a bond; that module is now the largest permissionless staking module on Ethereum by ETH staked.

In January 2026 the protocol launched Lido V3, adding stVaults: separate vaults with their own operator, fee and policy settings that can still mint stETH for liquidity. Permissionless vault minting followed in March 2026. The strategic direction is clear — Lido is positioning itself as staking infrastructure rather than a single pool — and it matters to depositors because the pooled product you use may be one of several, with different operators and different risk.

Who governs it

Lido is governed by LDO token holders: 1 billion maximum supply, with about 829.6 million circulating in October 2026. Token governance over a protocol holding billions in user deposits is an obvious conflict, and Lido addressed it directly. Dual Governance, enacted through LIP-28 and live since 30 June 2025, gives stETH holders a brake.

Under that system, stETH, wstETH and withdrawal NFTs can be locked in an escrow contract to signal opposition. At 1% of staked ether, a DAO proposal's execution is delayed by a dynamic timelock that grows with the amount locked, reaching 45 days at 10%. At 10%, the protocol enters a rage-quit state: no proposal can execute until the dissenting stakers have withdrawn. It is an unusually concrete answer to a real governance problem, and worth knowing about even if you never vote.

The risks that do not appear in the APR

Smart contract risk is first. A liquid staking token is a claim enforced by code, and the code holds other people's ether. Audits reduce this risk; they do not remove it.

Slashing risk is second, and it is socialised. If an operator is penalised for a provable protocol violation, the loss hits the pool and therefore every holder, not just the users who deposited with that operator. Correlated failures across many validators at once are the scenario that hurts.

Liquidity risk is third, and it is the one retail users underestimate. Your exit has two doors: the protocol's withdrawal queue and the secondary market. In a calm market both work. In a stressed market the first door is slow and the second door is cheap only for whoever gets there first.

Then there is the risk you create yourself. Because stETH is accepted as collateral, it can be borrowed against, and the borrowed ETH staked again. That loop amplifies yield and amplifies liquidation risk, and it turns a 6% token discount into a solvency event for the leveraged holder. If you lend your receipt token for extra yield, you have added a counterparty to a position that was already a claim on a pool. The mechanics are the same ones behind impermanent loss in liquidity pools, with more leverage on top.

Liquid staking versus the alternatives

Solo staking keeps custody and full rewards but demands 32 ETH, hardware and operational attention. Custodial staking outsources everything, including your keys. Liquid staking trades a fee and some protocol risk for liquidity and low minimums. Restaking, which reuses staked ether to secure additional services, adds another layer of reward and another layer of risk on top — restaking is best understood as a separate decision, not an upgrade.

None of these is universally better. The right question is what you need the capital to do. If the answer is nothing, plain staking is simpler and cheaper. If the answer is trade, collateralise or move across chains, liquid staking is the only structure that allows it.

What to check before you deposit

Read the fee and how it is split. Confirm the operator set and whether you can see per-operator performance. Check the withdrawal mechanics and whether there is a buffer for fast exits, because that is what separates a smooth exit from a week of waiting. Look at the token's secondary-market depth, not just its market cap. And check governance: who can change the fee, the operator set or the contracts, and what stops them.

Finally, size the position for a discount. If you might need the money in a hurry during a market panic, assume the token will trade below par exactly when you want to sell it, and assume the withdrawal queue will be longest at the same moment. Liquid staking works well when you treat it as a long position with a liquidity feature. It works badly when you treat the liquidity as guaranteed.

The bottom line

Liquid staking converts a locked yield into a tradable one, and that is a real improvement in capital efficiency. It does not make staking riskless or instant. The token is a share of a pooled stake, its price is a market quote, its exit has a queue, and its operators are other people. Understand those four things and the product makes sense; skip them and you are holding a receipt whose terms you never read.