Liquid staking lets you stake a proof-of-stake asset — most commonly ETH — and receive a second token in return that represents your staked position. That token, called a liquid staking token (LST), can be traded, lent, or posted as collateral while the underlying stake keeps earning rewards. Staking without giving up liquidity, in other words. That's the pitch, and it's mostly accurate.
The confusion around liquid staking usually comes from the token itself. People hold stETH or rETH and assume they hold staked ETH. They don't, quite. They hold a claim on a protocol that holds staked ETH — and the difference between those two things is exactly where the risks live. This post covers the mechanism: how the claim is created, why it trades close to the underlying, and what stacks on top of ordinary staking risk. If you're weighing liquid staking against running your own validator, that comparison is covered separately.
How Liquid Staking Actually Works
Start with the problem it solves. Staking on Ethereum means locking ETH to a validator — 32 ETH per validator, validator keys, uptime requirements, and an exit queue between you and your money. That's a real commitment of capital and attention, and proof of stake depends on enough people making it. Most holders won't. Liquid staking protocols exist to collect the capital from people who want the yield without the operation, and route it to operators who do the work.
The flow, using Ethereum as the example:
- You deposit ETH into the protocol's smart contract. No validator, no hardware, no 32 ETH minimum.
- The protocol pools deposits and assigns them to node operators. Operators run the actual validators. Your funds are batched with everyone else's into 32 ETH chunks; you never hold validator keys.
- The protocol mints an LST to your wallet. This is your receipt — a transferable token representing your share of the pooled stake.
- Rewards accrue to the token. As validators earn, your position grows.
- You exit one of two ways. Redeem through the protocol (burn the LST, wait for validators to exit through Ethereum's withdrawal queue), or just sell the LST on the open market and let someone else hold the claim.
Step four hides a design fork worth knowing about, because it changes how the token behaves in practice. Some LSTs rebase: your stETH balance literally increases day by day, one token targeting one ETH. Others are reward-bearing: your rETH quantity never changes, but each rETH becomes redeemable for progressively more ETH. Same economics underneath, different accounting — and the difference matters for DeFi integrations and, in many jurisdictions, for tax treatment. Neither model is "correct." The Lido vs Rocket Pool comparison walks through a live example of each.
Where the Peg Comes From
An LST trades close to the value of the ETH backing it, and it's worth understanding why, because the mechanism has a speed limit.
The anchor is redemption arbitrage. If stETH trades below its backing, anyone can buy it at the discount, redeem it through the protocol for the full amount of ETH, and pocket the difference. That buying pressure closes the gap. Since Ethereum's Shapella upgrade in April 2023 enabled staking withdrawals, this loop is real and continuously exercised — which is the main reason major LSTs now hold their peg tightly in normal conditions.
But notice the constraint: redemption runs through Ethereum's validator exit queue, and the consensus layer has no idea liquid staking exists. Ethereum sees validators, not LSTs. Exit throughput is rate-limited by protocol rules, which means the arbitrage closes gaps at the speed of the queue, not the speed of the market. In calm conditions that's invisible. Under stress — when many holders want out at once — the exit door narrows exactly when demand for it spikes, and the secondary market price can detach from backing until the queue catches up.
That's not a hypothetical. In June 2022, stETH traded roughly 5–7% below ETH for weeks as leveraged holders (Celsius, Three Arrows Capital) were forced to sell into a market with no redemption path at all — withdrawals didn't exist yet. The protocol functioned fine. The backing was intact. What broke was the assumption that a claim on locked ETH would always trade like ETH. Post-Shapella, that episode is harder to repeat, but the structural lesson stands: the peg is an arbitrage outcome, not a guarantee.
What Stacks on Top of Ordinary Staking Risk
Liquid staking doesn't remove staking risk — slashing, operator downtime, the base volatility of the asset. It adds layers on top:
Smart contract risk. Your claim lives in a contract. A bug in minting, accounting, or withdrawal logic is a direct loss vector that a solo staker simply doesn't have.
Socialized slashing. If an operator in the pool gets slashed, the loss is spread across all LST holders. You're exposed to the worst operator in the set, not the average one — mitigated by the protocol vetting operators and, increasingly, by requiring operators to post their own collateral first.
Depeg risk. Covered above. You can always sell an LST; you can't always sell it at par.
Concentration. This is the one with system-level consequences. Liquid staking pools stake far more effectively than individuals, so stake concentrates — at times, roughly three in ten staked ETH have sat under a single protocol's operator set, with the operator list controlled by that protocol's governance token. A protocol controlling a large fraction of validators is a governance and censorship surface for Ethereum itself, whatever the intentions of the people running it. This is the most persistent criticism of liquid staking, and it's a fair one.
What's Changing
Two structural shifts are worth watching. First, LSTs became the base collateral for restaking — reusing staked ETH to secure additional systems on top of Ethereum. That extends the liquid staking stack another layer, with its own risks, and it's the subject of the next post. Second, distributed validator technology (DVT) is being adopted inside liquid staking protocols: splitting a single validator's key across multiple machines and operators, so one operator's failure or misbehavior no longer maps one-to-one onto slashing events. It directly attacks the operator-risk and concentration problems, which is why the major protocols are integrating it.
On the concentration question itself: no protocol-level fix is scheduled. Proposals exist — from self-limiting (tried, voted down by the dominant protocol's own token holders) to changes in Ethereum's issuance curve — but as of mid-2026 this remains an open argument, not a roadmap item.
What Would Confirm This Direction
Redemption queues functioning under a genuine stress event — a large depeg closing via arbitrage rather than intervention — would be the strongest confirmation that the post-Shapella design holds. Beyond that: DVT adoption spreading across major protocols, operator sets growing more numerous and independent, and LST share of total stake growing without any single protocol's share growing with it.
What Would Break It
A major LST smart-contract failure with losses reaching holders. A sustained depeg while redemptions are functioning — that would mean the market is pricing a risk the mechanism can't see, which is worse than a queue delay. Regulatory action classifying LSTs as securities in major markets, cutting them off from the DeFi integrations that justify their existence. And on the systemic side: a single protocol's share of validators grinding upward past the point where Ethereum's neutrality credibly depends on one governance token.
Timing
Now: Liquid staking is live, mature, and one of the largest categories in DeFi by locked value; withdrawals and redemption arbitrage are battle-tested in calm conditions, less so under severe stress. Next: DVT integration and the restaking stack maturing on top of LSTs — that's where new risk is accumulating. Later: any protocol-level response to stake concentration; the arguments are active, the code is not.
Boundary
This covers what liquid staking is and how the claim mechanism works. It is not a yield comparison, an assessment of any protocol as an investment, or a recommendation to stake in any form. Which protocols' concentration levels matter, and at what thresholds — those are tracking questions. This is the static explanation; the tracked version lives elsewhere.
