Staking locks capital. Liquid staking gives you a receipt for staked capital so you can keep using it. That receipt is the entire invention, and it is also the entire risk: from the moment the receipt exists, two things are trading in the market — the underlying stake and the claim on it — and they are not always the same price.
What the receipt actually is
Deposit ETH with a liquid staking protocol and you receive a token representing a claim on that stake plus accumulated rewards. Lido's stETH rebases, so your balance grows and one stETH stays worth roughly one ETH. Rocket Pool's rETH appreciates instead, so your balance stays fixed and each token is worth more over time. Coinbase's cbETH is the institutional-facing version of the same idea; sfrxETH from Frax is one of several where a higher advertised yield is the first clue that something additional is being taken on.
Both designs work. The rebasing token is easier to reason about in DeFi integrations and harder to use as collateral in systems that do not handle changing balances. The appreciating token is the reverse. Neither is better in the abstract, and picking one for a lending market means understanding which convention the market expects.
Where the peg breaks
A liquid staking token trades at a discount when more people want out than the available exit liquidity allows — during a market shock, when withdrawals are queued, or when a specific protocol is under suspicion. In normal conditions the discount is a few basis points and nobody notices. In a stress event it can be several percent, and it recovers on a timescale nobody can promise.
Ethereum's withdrawal queue is the mechanism behind that: exits are processed at a rate limit, so a rush of withdrawals creates a backlog measured in days or weeks. Anyone holding the receipt as collateral at a loan-to-value that assumed 1:1 is now exposed to a discount and a liquidation at the same time.
The rest of the risk surface
Smart contract risk is higher than it looks, because these protocols interact with validator clients, withdrawal credentials, and in some designs upgradeable components. Centralisation is the systemic worry: when one protocol holds a large share of staked ETH, its validator choices and its governance affect the network itself, not just its own users. Slashing passes through to holders in most designs. And the operator layer matters — protocols that curate a limited set of node operators are trusting those operators to behave.
Why restaking changes the shape
Restaking lets the same staked asset secure additional services, which pays more and layers new slashing conditions on top of the original ones. The result is leverage on security: the same capital, more yield, and more ways to be penalised.
The question to ask is not whether the extra yield is worth it, but whether the risks are correlated. If the restaked services are exposed to the same market event as the underlying stake, the diversification is illusory and the position is simply leveraged.
How to choose without being naive
Judge a protocol on its operator set and how it is chosen, the audit history including what was found and left unfixed, the depth of secondary liquidity for the receipt, whether withdrawals have ever been tested under stress, and governance — who can change the contracts and how quickly.
Then decide what you are actually buying. Hold it as a yield-bearing asset and it is a straightforward trade. Loop it through a lending market and you have a directional leveraged position with a depeg risk attached. Both are legitimate; only one of them is what most people think they are doing.
Next: how the staked assets underneath this market are secured and rewarded. Join our Discord.
References
Next, read Fiat-Backed Stables 101. More in DeFi Markets.