Staked ETH already earns a base yield, currently around 3% to 4%. Restaking lets that same staked capital secure additional networks and earn a second time. Liquid restaking yield is the return that comes from this layered model, packaged into a single token you can hold, trade, or use elsewhere in DeFi.

The pitch is easy to like. The reality deserves more scrutiny. Much of the headline yield in 2026 comes from token emissions rather than durable fee revenue, and the risk stack runs deeper than standard staking. For an allocator optimizing an ETH position, the question is not whether the extra yield exists. It does. The question is whether the incremental return compensates for the incremental risk, and how large a position makes sense.

This guide breaks down how liquid restaking works, where the yield actually comes from, the specific risks that matter for institutional capital, and a framework for sizing an allocation. The goal is a clear read on a strategy that sits a notch above plain staking on the risk curve.

What Is Liquid Restaking, and Why Does It Exist?

Restaking takes ETH that already secures Ethereum and extends its economic security to other protocols. Those protocols, called Actively Validated Services (AVSs), include bridges, oracle networks, and data availability layers that need security but do not want to build their own validator set from scratch. In return for accepting additional slashing conditions, restakers earn extra rewards.

Think of it as collateral reuse. In traditional finance, rehypothecation lets the same collateral support more than one obligation. Restaking applies a similar idea to staked ETH: one asset, multiple security commitments, multiple reward streams. EigenLayer pioneered the model and still holds the dominant share of the market.

What does a liquid restaking token actually represent?

A liquid restaking token (LRT) is a claim on your underlying ETH plus the rewards accruing from base staking and any AVSs your stake helps secure. Protocols like Ether.fi, Renzo, Kelp DAO, and Puffer handle the operator selection and AVS allocation, then issue you a token (eETH, ezETH, rsETH, pufETH) that stays liquid while the position earns in the background.

Four parties make the system work: the staker who provides capital, the LRT protocol that pools deposits and abstracts the complexity, the node operator that runs the infrastructure, and the AVS that purchases security. Your token sits on top of all of them, which is the source of both its convenience and its risk. Understanding that structure is the first step in pricing what you are actually holding.

Where Does Liquid Restaking Yield Actually Come From?

This is the question most worth getting right, because the answer shapes how durable your return is.

The three-layer yield stack

Liquid restaking yield builds in three layers. The base is Ethereum staking yield, currently around 3% to 4%. On top of that sit AVS rewards, which historically have added roughly 1% to 2% depending on which services your stake secures and how their reward structures perform. The third layer is points and token incentives, which vary widely and can inflate a quoted APY well beyond the real, recurring return.

Add the durable layers together and most LRT positions produce something in the 4% to 7% range in real terms. Headline numbers run higher when speculative incentives are counted, and looped positions can reach 8% to 15% or more, with liquidation risk attached. So a quoted "12% APY" might decompose into roughly 3% base staking, 1% to 2% of AVS rewards, and the rest in emissions and points that may not persist. If you are considering the looped route, the mechanics and the failure modes are worth studying first; our breakdown of leverage looping in DeFi covers how those loops behave under stress.

Emissions versus fee revenue, the distinction that matters

Here is the part allocators miss. Most restaking yield in 2026 still comes from EIGEN token emissions, not from fees that AVSs pay for security. Emissions are a legitimate way to bootstrap a network, but they are inflationary by nature and compress as the token price adjusts and as the protocol gradually shifts toward fee-based rewards. AVS fee revenue has not yet grown to the scale that current total value locked implies.

For a treasury or fund, that means a quoted APY can be real today and materially lower in a year. The recurring, fee-based portion is the number that belongs in a forward-looking allocation model; the emissions portion is closer to a temporary incentive. Separating the two is the same discipline you would apply to any structured product, and it is the core idea behind evaluating a yield platform beyond its headline APY.

The LRT Market in 2026: Ether.fi, Renzo, Kelp, Puffer

EigenLayer remains the dominant restaking protocol, holding the large majority of the market. Its total value locked has settled into the high-single to low-double-digit billions through 2026, down from a 2025 peak above $20 billion, as the sector moved from rapid growth into a consolidation phase. Plain liquid staking, led by Lido, still commands more capital than pure restaking, which tells you something about where the broader market currently places its trust.

The major LRT providers differentiate on design rather than headline yield. Ether.fi is the largest by deposits and issues eETH, with a wrapped version used widely as DeFi collateral. Renzo (ezETH) and Kelp DAO (rsETH) each take their own approach to operator selection and how they spread stake across AVSs. Puffer (pufETH) markets anti-slashing technology intended to reduce validator penalties at the infrastructure level. These choices change the risk profile more than the return, so the provider you pick is itself a risk decision rather than a yield decision.

Two design questions separate the providers in practice. The first is operator concentration: a token spread across many vetted operators behaves differently in a slashing event than one concentrated in a few. The second is AVS selection, since more experimental services can carry richer rewards and looser slashing conditions, while conservative allocations trade some yield for a tighter risk profile. Neither choice shows up in a quoted APY, which is why reading a provider's allocation policy matters more than comparing top-line numbers.

Restaking is not only an Ethereum story. On Solana, Jito has built restaking on top of its liquid staking position, offering base staking, MEV rewards, and rewards from Node Consensus Networks. Bitcoin restaking, through protocols such as Babylon, has grown from a standing start into a multi-billion-dollar category. The primitive is spreading across chains, but Ethereum remains where most institutional restaking capital sits today.

What Are the Real Risks of Liquid Restaking?

Liquid restaking carries more risk than plain ETH staking because each layer adds a new way to lose money. For institutional capital, three categories deserve specific attention.

The slashing stack

Slashing is live, not theoretical. Restaking risks are additive: a validator can be penalized on Ethereum for downtime or double-signing, and each AVS your stake secures adds its own slashing conditions on top. If several AVS networks fail at once, whether through bugs, attacks, or network partitions, an operator can face penalties across multiple protocols simultaneously. This correlated, or cascading, slashing is the scenario treasury managers worry about most, and it is the reason operator and AVS selection matter more than the quoted yield.

EigenLayer has introduced mitigations, including opt-in slashing, exclusive stake allocation through Operator Sets, and clearer separation of slashable exposure. Proposed operator safety scores and AVS risk ratings would help allocators price this risk directly, but as of mid-2026 those tools are still maturing rather than fully deployed.

Depeg, liquidity, and withdrawal-queue risk

An LRT can trade below its ETH-equivalent value during market stress. A 5% to 10% depeg is not unusual when liquidations cascade through DeFi, and if you have used the token as leverage collateral, even a modest depeg can trigger forced selling. Exits are not instant either. Depending on the route in, unwinding a position can take days or weeks, and secondary-market liquidity can thin out exactly when you most want to leave.

For an institution, this is a duration and liquidity-management problem as much as a yield decision. One way to reduce the underlying ETH price exposure is to pair the position with a hedge, an approach we cover in our guide to delta-neutral strategies.

Layered smart contract risk

Holding an LRT means depending on the token contract, the EigenLayer contracts, and the contracts of every AVS in the delegation path. A bug anywhere in that chain can cause losses. The Kelp DAO incident in April 2026, an exploit of roughly $293 million at the bridge layer, is a concrete reminder that the failure point is often somewhere holders never examined. Diversifying across providers and reading which AVSs a token delegates to are basic precautions, not optional extras.

How Should an Institutional Allocator Approach LRT Yield?

Start with sizing. Because the risk stack is layered and slashing is live, a sensible position is one where a worst-case loss does not damage the broader portfolio. A treasury holding ETH might cap restaking exposure at a small share of its ETH sleeve and keep the remainder in plain staking or cash-like onchain instruments. Restaking is an incremental yield enhancement, not a core holding.

Then do the diligence the headline APY hides. Identify which AVSs the token secures and what their slashing conditions are. Separate the recurring, fee-based portion of the yield from token emissions, and assume the emissions component compresses over time. Check the provider's operator set and how stake is distributed, since concentration is what turns an isolated slashing event into a portfolio problem. If you already run an ETH yield program, restaking can sit alongside it; our overview of strategies for earning yield on ETH is a useful starting point for deciding where it fits.

A practical way to frame the decision is to model two yield numbers side by side: the durable, fee-based return you would still expect a year from now, and the headline figure that includes points and emissions. If the durable number alone clears your hurdle rate after accounting for slashing and liquidity risk, the position can stand on its own. If it only clears with emissions included, you are underwriting a temporary incentive, which is a different and shorter-dated bet.

Hedging and leverage deserve a deliberate decision rather than a default. Looping an LRT through lending markets raises the yield and the liquidation risk together, so it belongs only in mandates that can monitor and manage that exposure actively. Where the goal is capital preservation with a modest yield premium, an unleveraged and hedged position is the more defensible structure.

Regulation is moving toward clarity. The SEC's August 2025 policy statement on liquid staking signaled that non-managerial staking functions may avoid securities classification, which removed some uncertainty for US allocators, though restaking adds wrinkles that the statement did not fully address. Mapping where restaking sits on your risk spectrum, against stablecoin and lending strategies, is easier with a structured framework like our risk-tiered approach to crypto yield.

Conclusion

Liquid restaking yield is real, and for holders already committed to ETH, the incremental return can be attractive. The caveats are equally real. Much of today's yield comes from token emissions rather than durable fee revenue, slashing is live and additive across AVSs, and the convenience of a single liquid token rests on several layers that can each fail. Treat it as a measured yield enhancement, size it accordingly, and diligence the operators and AVSs behind any token you hold.

Platforms like Lucidly Finance are built for exactly this kind of decision, combining curated access with institutional-grade risk frameworks so you can capture onchain yield without managing every operator and AVS relationship by hand. If restaking earns a place in your allocation, let it be a sized, deliberate one.