Imagine you deposited ETH into a liquid restaking token at the peak of the 2024-2025 points cycle. Your dashboard showed double-digit APY figures, airdrop allocations were accumulating, and every protocol blog post described the new primitive as "ETH but working harder." By mid-2026 the EIGEN airdrop has run its course, LRT governance tokens have seen significant post-launch price decay, and the April 2026 Kelp exploit — roughly $292M of rsETH drained via a LayerZero bridge vulnerability — triggered billions of dollars in sector-wide withdrawals. The question is no longer whether restaking can produce yield. It is whether the incremental yield above plain Ethereum staking is fair compensation for a genuinely stacked set of risks.
The Honest Yield Stack: What Restaking Actually Pays in 2026
Base Ethereum validator yield sits at approximately 2.5–3% APY in 2026. With roughly 28% of all ETH staked on the network, issuance per validator is mechanically suppressed — this number is the floor beneath every restaking comparison. If you are holding a liquid restaking token, every percentage point of yield above this floor is the premium you are earning, and the price you are paying in additional risk.
A single, unleveraged, diversified LRT — weETH, ezETH, or rsETH — has historically delivered around 5–8% APY in ordinary market conditions once AVS rewards, DeFi incentives, and governance-token emissions are combined. That is a gross premium of roughly 2–5 percentage points over plain staking. Not nothing, but not dramatic either.
The higher 8–15% figures cited by aggregator dashboards typically include DeFi leverage loops and protocol-specific incentives that are not the LRT yield itself. Strip those out and the honest restaking premium narrows considerably. AVS rewards are denominated in a mix of ETH and stablecoin fees from fee-paying AVSs, plus EIGEN emissions and AVS governance tokens; the ETH-denominated portion is modest. EigenLayer's Incentives Committee, formed around the start of 2026, now directs rewards toward fee-paying AVSs rather than all validators, which is the right long-term direction — but it reduces passive yield today.
The honest forward yield for a passively held, unleveraged LRT in 2026 is likely in the mid-single-digit percentage range — meaningfully above plain staking, but subject to further compression as more AVSs transition to real fee models.
The Airdrop Era Is Over: What That Means for Forward Yields
The broad EIGEN airdrop distributions ran their course by 2025. No further broad distribution is scheduled, ending the "farm points, claim tokens" cycle that drove explosive LRT TVL growth from 2024 onwards. Restaking-sector TVL peaked in early 2026 before the April Kelp contagion triggered billions in sector-wide withdrawals; cross-protocol restaked ETH was still cited in the high-teens of billions of dollars afterward, though figures differ by measurement methodology.
With the speculative premium gone, forward LRT yield now depends almost entirely on genuine AVS fee revenue. LRT governance tokens — ether.fi's ETHFI, Renzo's REZ, Kelp's KEP — have seen significant post-launch price decay, reducing the fiat-equivalent yield for holders who mark governance-token rewards to market.
This shift matters for anyone sizing a restaking position today. Historical APY figures from the points era are not a reliable guide to what you will earn going forward. The honest forward expectation is mid-single-digit APY with a risk profile that plain staking simply does not carry.
Risks to Understand Before You Participate
Smart-contract risk is stacked in a way that is qualitatively different from ordinary DeFi. Your ETH passes through at least three independent contract layers: the LRT protocol itself, EigenLayer or EigenCloud core contracts, and each AVS's individual slashing conditions. A vulnerability at any layer can cause permanent loss of principal — not a temporary de-peg you can wait out, but irrecoverable loss. The April 2026 Kelp event, where a bridge-layer vulnerability (not a core slashing event) drained roughly $292M of rsETH, is the clearest demonstration that this risk is not theoretical.
Slashing surfaces multiply. Restaked ETH is subject to Ethereum's base slashing rules and every AVS's independent slashing conditions simultaneously. An operator misbehaviour on a single AVS can trigger losses across a user's entire restaked position. For a primer on how Ethereum validators earn and get slashed, the base-layer mechanics are worth understanding before adding AVS exposure on top.
De-peg and withdrawal-queue risk are structurally linked. LRTs trade on secondary markets and can diverge several percent from underlying ETH value during stress events. Withdrawing to native ETH requires queuing through both the LRT protocol and Ethereum's validator exit queue, which can take days to weeks. The mismatch between instant DeFi liquidity and slow on-chain exits is a known systemic fragility — and it becomes acute exactly when you most want to exit.
Leverage amplifies every one of these failure modes. Using an LRT as DeFi collateral to borrow and loop back into more restaking can generate double-digit yields, but it also means a modest de-peg triggers liquidations before any withdrawal is possible.
Contagion is real, as the broader sector learned in April 2026. The tight composability of restaking protocols means composability means one protocol's crisis is everyone's liquidity event — even for users who held no Kelp position directly. Oracle risk layers on top: LRT-collateralised lending protocols depend on price oracles, and an oracle that lags or misprices an LRT during a stress event can either accelerate liquidations or fail to flag genuinely impaired collateral.
A Decision Framework: Who Should Restake, Who Should Not
The restaking question is not binary. Different user types face materially different risk-reward profiles.
The conservative ETH holder — multi-year horizon, capital preservation as the priority — is best served by plain native staking or a battle-tested LST like stETH or rETH. The extra few percentage points from restaking do not justify the tail risk of a contract-level principal-loss event. For anyone curious about staking and earning yield inside Zelcore, the platform deliberately scopes its Earn feature to native delegated staking rather than restaking or leverage, which reflects this risk calculus.
The yield-seeker comfortable with smart-contract risk — already active in DeFi, understands approval hygiene, monitors positions regularly — is a reasonable candidate for a single, diversified, unleveraged LRT held in self-custody. Not deposited on an exchange, not used as loan collateral.
The DeFi power user who understands oracle risk and leverage-loop mechanics and actively monitors liquidation thresholds can access higher restaking yields, but should understand they are no longer earning a "staking yield." They are running a structured credit position with multiple simultaneous failure modes.
Three user types who should not restake: anyone who cannot explain what a slashing condition is; anyone who might need liquidity within a week without warning; and anyone who would only hold LRTs on a centralised exchange rather than in self-custody.
Self-Custody Is Non-Negotiable for Restaking Participants
An LRT held on a centralised exchange is a custodial IOU. The exchange holds the token; you hold a claim against the exchange. Both the restaking yield and the underlying ETH are exposed to exchange counterparty risk on top of all the protocol risks already described. The principle here is the same as for any digital asset: your keys, your coins. Hold weETH, rsETH, or any LRT in your own non-custodial wallet so that your relationship is with the smart contract, not an intermediary.
Understand the withdrawal queue before you need it. LRT redemptions to ETH are not instant; during stress events the queue lengthens and secondary-market de-pegs can make selling the LRT the faster but more costly exit. Be cautious about using LRTs as DeFi collateral: if the de-peg event that triggers your liquidation is the same event making you want to exit, you cannot withdraw fast enough to protect yourself. Zelcore holds weETH and other ERC-20 LRTs natively under your own recovery phrase — the self-custody principle for LRTs is identical to the principle for plain ETH.
Pre-commit checklist before entering a restaking position:
- Can you name every smart-contract layer your ETH passes through?
- Do you understand the withdrawal-queue mechanics for your chosen LRT?
- Are you holding the LRT in your own wallet under your own keys?
- Is the position sized so a total loss would not materially harm you?
- Have you read the April 2026 Kelp post-mortem?
Series Close: Restaking Is Real Infrastructure, Yield Is a Risk Premium
This is the fifth and final part of the series. Part 1, restaking vs staking from first principles, established the core mental model: restaking layers additional security obligations and rewards on top of staked ETH, and the intuitions of plain staking do not transfer. Part 2 examined EigenLayer's AVS architecture — genuine, valuable security infrastructure, but with AVS fee sustainability still being proven at scale and token emissions running down. Part 3 mapped the LRT landscape across weETH, ezETH, and rsETH, each representing a different operator strategy, fee structure, and risk profile; diversification across LRTs reduces but does not eliminate protocol-specific risk. Part 4 documented the April 2026 Kelp contagion in detail — a bridge-layer vulnerability that proved the risk is not theoretical and that sector-wide withdrawal pressure can materialise from a single protocol failure.
The conclusion is straightforward: restaking is functioning cryptoeconomic infrastructure that provides real, if modest, yield above plain staking. But the premium is payment for genuine, stacked, compounding risk that plain staking does not carry. Size your restaking position accordingly, hold it in self-custody, avoid leverage, and know your exit before you enter.
Key Takeaways
- The honest unleveraged LRT yield in 2026 is likely mid-single-digit APY — roughly 2–5 percentage points above plain Ethereum staking, not the double-digit figures of the points era.
- The airdrop and points cycle is over; forward yield depends on genuine AVS fee revenue, which is still maturing.
- Smart-contract, slashing, de-peg, withdrawal-queue, and contagion risks are all real and compound each other; one exploit at any layer of the stack can cause permanent principal loss.
- Conservative holders should stick to native staking or proven LSTs; only users who actively monitor DeFi positions and understand liquidation mechanics should consider LRT exposure.
- Self-custody is mandatory — an LRT on an exchange adds counterparty risk to an already complex stack, with no corresponding yield benefit.
