Liquid Restaking Tokens Compared 2026: eETH, ezETH, rsETH, pufETH, weETH

If you staked ETH in 2022, the hardest decision was "Lido or Rocket Pool." In 2026 there are over a dozen liquid restaking tokens competing for the same deposit, each pointing at a slightly different cocktail of EigenLayer AVSs, Symbiotic networks, operator sets, and insurance mechanisms — and each with its own depeg story, points scheme, and real underlying yield.

The pitch for LRTs hasn't changed: stake ETH, earn staking yield, earn restaking yield on top, get a liquid token you can use elsewhere in DeFi. The reality is that the top LRTs now differ in ways that matter — slashing exposure, underlying AVS mix, redemption mechanics, and whether the "yield" you see in DeFi dashboards is real revenue or just points inflation.

This guide compares the biggest LRTs honestly, with the specifics you need to decide which one — if any — belongs in your portfolio.

TL;DR

• LRTs wrap restaked ETH into a transferable token that earns base staking yield (~3%) plus restaking rewards from EigenLayer AVSs, Symbiotic networks, or both
• The five dominant LRTs in early 2026 are eETH/weETH (Ether.fi), ezETH (Renzo), rsETH (Kelp), pufETH (Puffer), and rswETH (Swell) — with Ether.fi the clear market leader
• Real restaking yield has compressed from double-digit "points-era" expectations to roughly 1–3% on top of base staking, making LRT choice more about risk profile than raw APY
• Slashing risk is real — the specific AVSs an LRT opts into determine what can slash your principal, and most protocols still publish this only partially
• Depeg events happened in 2024–2025 as withdrawal queues lagged market stress; direct redemption mechanics now matter more than secondary market depth
• The practical choice reduces to: Ether.fi for size and distribution, Puffer for anti-slashing design, Kelp for DV (distributed validator) positioning, Renzo for multi-chain reach, Swell for EigenLayer native integration

What an LRT Actually Is

A liquid restaking token is a receipt token issued by a protocol that:

1. Takes your ETH (or wETH, or an LST like stETH)
2. Stakes it with Ethereum validators
3. Restakes the validator credentials into EigenLayer, Symbiotic, or both — meaning the same capital also secures other services and earns fees
4. Gives you a token representing your share of the total pool

That token is transferable, composable across DeFi, and accrues value as both staking rewards and restaking rewards flow into the pool. It is also subject to the combined slashing risks of everything the protocol opted into.

The key distinction from plain LSTs (stETH, rETH):

The extra 50–300 bps is real yield today. It is not the 20%+ APYs that points-era dashboards were implying. Anyone still quoting those numbers is either lying or including points valuations that have since deflated.

The Five LRTs That Matter

Ether.fi — eETH / weETH

Size: Largest LRT by TVL through 2025 and into 2026. In the $6–9B range depending on market conditions.

Design:

• Users hold eETH (rebasing) or weETH (non-rebasing wrapper, preferred for DeFi)
• Uses its own node operator set for staking with DV (distributed validator) technology via SSV Network
• Opts into a curated set of EigenLayer AVSs plus selected Symbiotic vaults
• Stake-key separation: withdrawal credentials held by Ether.fi, validator keys held by node operators — reduces single-point-of-failure slashing

Yield sources:

• ~3% ETH staking
• ~1–2% EigenLayer AVS fees
• ETHFI token incentives (variable)

Why it dominates:

• First mover with clean product and aggressive distribution
• Deepest DeFi integrations (used as collateral in Aave, Morpho, Spark, Pendle)
• Active CeDeFi partnerships — Bybit, Binance Earn integrations pull real balance-sheet demand

Risks to name:

• Largest concentration of AVS exposure — a slashing event on any opted-in AVS could impact pool
• ETHFI emission schedule affects reported APY vs real yield
• Management-layer centralization (Ether.fi team curates AVS list)

Who it suits: Anyone who wants the most liquid, most composable LRT and can accept a diversified AVS slashing profile.

Renzo — ezETH

Size: Top 3 LRT by TVL. Heavy multi-chain presence.

Design:

• ezETH is non-rebasing from day one (simpler DeFi composability)
• Natively bridged to multiple L2s and alt-L1s via Connext/Everclear
• Operates with a diversified operator set and tiered AVS exposure
• REZ token governs protocol decisions and AVS selection

Yield sources:

• ~3% ETH staking
• EigenLayer AVS fees
• REZ incentives

Why it stands out:

• Deepest L2 and cross-chain distribution of any LRT — ezETH is native on Arbitrum, Base, Blast, Linea, BNB
• Strong DEX liquidity across chains makes it the most tradable LRT under stress
• Clean non-rebasing design — what you hold is what you get, no accounting gotchas

Risks to name:

• Experienced a meaningful depeg event in April 2024 that still shapes how people think about LRT redemption queues
• Cross-chain distribution is a double-edged sword: more liquidity venues, more bridge risk
• Reliance on external operators for restaking decisions

Who it suits: Users who want ETH restaking exposure on L2 without bridging ETH back to mainnet, and who prioritize tradability over anti-slashing architecture.

Kelp DAO — rsETH

Size: Top 5 LRT. Focus on institutional and DV distribution.

Design:

• rsETH is non-rebasing
• Heavy use of distributed validator technology (DV) through Obol and SSV
• Curated AVS selection with conservative opt-in policy
• KEP points converted into KELP token at TGE; governance active

Yield sources:

• ~3% ETH staking
• Restaking rewards, weighted toward lower-slashing AVSs
• KELP incentives

Why it stands out:

• DV-first architecture means validator duties are spread across multiple nodes, reducing correlated slashing risk
• Conservative AVS opt-in is deliberately less aggressive on APY in exchange for cleaner risk
• Strong in ETH2 institutional channels and LSDFi integrations

Risks to name:

• Lower headline yield than peers — which is the point, but dashboards don't always show that
• Smaller DeFi footprint than eETH/ezETH — fewer places to deploy rsETH
• Dependence on Obol/SSV for DV infrastructure adds an external-protocol assumption

Who it suits: Users prioritizing lower correlated-slashing risk and who don't need the absolute highest APY.

Puffer — pufETH

Size: Top 5–7 LRT. Differentiated positioning around anti-slashing.

Design:

• pufETH is non-rebasing
• Core differentiator: Secure-Signer and anti-slashing hardware requirements for node operators
• Low entry validator bond lets solo operators participate — more decentralized operator set
• Native integration with EigenLayer AVSs and its own operator incentive layer

Yield sources:

• ~3% ETH staking
• EigenLayer AVS fees
• PUFI token (launched 2024)

Why it stands out:

• The most explicit pitch on protecting principal from slashing — every node runs hardware-enforced slashing protection
• Aligns with Ethereum's solo-staker ethos
• Clean, non-rebasing design with active DeFi integrations

Risks to name:

• Hardware requirements limit how fast the operator set can grow
• Smaller TVL means shallower secondary liquidity in stress
• Most of the anti-slashing claim is a promise about correlated risk, not a guarantee about any single AVS

Who it suits: Users who care most about minimizing slashing blast radius and who want to support Ethereum's decentralized operator culture.

Swell — rswETH / swETH

Size: Top 7–10 LRT. Tightly integrated with EigenLayer.

Design:

• rswETH is the restaked variant; swETH is the pure LST
• Direct, deep EigenLayer integration with opt-in AVS selection
• SWELL token governance
• Swell L2 (rollup) provides downstream product surface for rswETH composability

Yield sources:

• ~3% ETH staking
• EigenLayer AVS fees
• SWELL incentives plus Swell L2 ecosystem rewards

Why it stands out:

• One of the earliest and deepest EigenLayer-aligned LRTs
• Own L2 creates a captive ecosystem for rswETH usage
• Clear split between LST (swETH) and LRT (rswETH) lets users choose exposure

Risks to name:

• L2 ecosystem ambition adds execution risk beyond just running a staking protocol
• Smaller TVL than Ether.fi/Renzo limits DeFi depth
• Heavy concentration in EigenLayer as the restaking venue

Who it suits: Users who want maximum EigenLayer exposure and are comfortable with a protocol that's building a broader L2 ecosystem.

What Happened to the Points Economy

Through 2024, LRT headline APYs were dominated by points — off-chain loyalty systems that promised future token airdrops. Pendle made the points tradable, leveraged farmers stacked 20–30x multiplier positions, and dashboards showed "implied yields" well north of 20%.

Through 2025 and into 2026, almost all those points campaigns converted into token unlocks. The token prices went down. The actual restaking yield was always closer to 1–3% on top of base staking. Points didn't create yield — they borrowed it from future token issuance.

The practical consequence: when you evaluate an LRT in 2026, ignore any dashboard that still prices points as yield. Look at:

• Base ETH staking yield (~3%, roughly the same everywhere)
• Restaking AVS fees (reported gross, net, or both — check which)
• Token emissions (real but dilutive)
• Any CeFi integration premium

If the three of those add up to 4–6%, that's typical LRT reality today. Anything meaningfully above that is either a short-lived incentive or a risk you're not seeing.

Slashing — The Part Dashboards Don't Show

Every LRT's marketing emphasizes yield. The honest framing is that you take on extra slashing risk for that yield.

On vanilla Ethereum, slashing is well-understood: behave badly, lose 1 ETH up to 100% of stake, penalties bounded by Casper rules. On EigenLayer or Symbiotic, slashing is defined per AVS — each service sets its own slashing conditions, and opted-in validators can be penalized for violations of those conditions.

What matters for you as an LRT holder:

• What AVSs has your LRT opted into? Many LRTs list this partially. Press for the full list
• What's the maximum slashing per AVS? Some AVSs cap slashing at small percentages; others allow larger penalties
• Is slashing socialized across the LRT pool or concentrated on specific operators? In most LRT designs, it's socialized — which means even conservative users eat losses from any slashed operator in the pool

No real slashing event has hit a major LRT's principal as of early 2026. That is the point — the test hasn't come yet. Everyone involved is running an untested insurance product at scale.

Depeg Risk and Redemption Mechanics

An LRT can trade below its "fair" NAV on secondary markets even if nothing fundamental is wrong, if:

• The native withdrawal queue is slow
• Leveraged positions unwind faster than redemptions clear
• A large holder needs liquidity immediately

April 2024's ezETH depeg was the clearest example — a points-campaign-end trigger moved ezETH several percent below NAV for hours, liquidating leveraged positions on Morpho and elsewhere.

The structural fix most LRTs adopted:

• Faster native redemption paths (hours, not days, for non-stressed conditions)
• Arbitrage-friendly mint/burn mechanics to pull secondary price back to NAV
• Cleaner accounting for withdrawal queue priority

Practical guidance:

• Don't lever LRTs heavily against volatile collateral — 5x looping in Morpho is not free money
• Hold the LRT variant your target DeFi venue prefers (weETH > eETH for Aave/Morpho)
• Know your exit path — if you'd need to exit 10% of your position in an hour, what venues support that?

Which LRT Fits Which Strategy

Passive hold, maximum composability → weETH (Ether.fi). Deepest DeFi integration, most battle-tested distribution.

L2-native yield → ezETH (Renzo). Native on almost every major L2 without round-tripping to mainnet.

Conservative slashing exposure → rsETH (Kelp) or pufETH (Puffer). Lower headline yield, deliberately tighter risk.

EigenLayer-aligned positioning → rswETH (Swell). If your thesis is specifically that EigenLayer captures the restaking economy, Swell is the cleanest direct expression.

Diversified LRT basket → all five in size-weighted chunks. Slashing risk is not perfectly correlated; a basket smooths the tail.

What to Watch Through 2026

• First real slashing event — the day a major AVS actually slashes validators at scale will be the day the market learns what these LRTs are really worth
• Symbiotic vs EigenLayer market share — most LRTs now opt into both; which restaking layer wins share matters for fee distribution
• Institutional LRT products — BlackRock-style wrappers and ETF proposals around restaking are in early stages
• DV adoption — distributed validator tech (Obol, SSV) is the strongest structural answer to correlated slashing; protocols that adopt it should outperform on risk-adjusted basis
• Yield compression vs AVS growth — if new high-value AVSs come online, real yields rise; if not, LRT yields converge with pure LST yields and the trade flattens

LRTs are neither the 20%-APY miracle of 2024 nor a scam. They are a real, slightly-higher-yield alternative to plain ETH staking with a different and still-evolving risk profile. In 2026, picking the right one is less about chasing APY and more about which slashing model, which operator set, and which redemption mechanic you actually trust.

More on the underlying stack: our EigenLayer Restaking Guide explains the primary layer, and Symbiotic vs EigenLayer covers the alternative restaking venue most of these LRTs also use.