Restaking Explained 2026: EigenLayer, Slashing, and What Changed After weETH

Restaking lets already staked capital secure extra services. You collect a premium. You also accept extra ways to lose principal.

That idea still holds in August 2026. The product around it does not. Points farming faded. Slashing went live. A major liquid restaking wrapper was exploited. ether.fi then split restaking out of weETH.

This guide explains the market as it exists on 28 August 2026. Price and TVL lines below are dated snapshots. It covers EigenLayer, EigenCloud, Symbiotic, Babylon, Jito, and the liquid tokens built on top. It also covers when you should stay in plain staking.

Most guides still ranking for this topic are quoting February and May TVL figures as if they were current, and describing weETH the old way. That’s your first signal to distrust a restaking article with no visible date on it.

Key takeaways for 2026

• Restaking is shared security, not free yield.
• EigenLayer is the security marketplace. EigenCloud is the developer platform built on it.
• Live slashing is isolated by Operator Set. That limits one service from draining every allocation at once.
• The costly 2026 failure was a bridge and lending stack, not an EigenLayer slash.
• weETH is no longer a bundled restaking token. Restaking now sits in a separate ether.fi product on Symbiotic.
• Stacked ETH yields often land near 4% to 8%, not the old 20% screenshots.
• Bitcoin restaking through Babylon is a separate category. It is not an Ethereum clone.
• Protocol usage and the EIGEN token are different stories. EIGEN traded near $0.20 on 28 August 2026, far below its $5.65 high.
• When ether.fi moved restaking into a separate ticker, reported weETHs demand sat near half a percent of the staking book. Bundled restaking was a default, not a choice.

As of 28 August 2026

July TVL prints are labeled July because that is the last clean comparison set used in this draft. Refresh them from DefiLlama before you hit publish.

What restaking is, and what it stopped being

Restaking reuses staked assets as collateral for more than one job.

On Ethereum, your ETH already secures the base chain. Restaking lets the same ETH help secure extra services. Those services are usually called Actively Validated Services, or AVSs.

You can think of it as insurance underwriting. Ethereum’s economic security is the underwriter. The AVS is the insured protocol. The extra yield is the premium. Slashing is the claim.

Three layers get mixed together in most explainers. Keep them separate.

1. The marketplace that takes collateral and assigns it to services.
2. The liquid token that wraps that position so you can trade it.
3. The service that buys the security, such as a data layer or oracle.

Restaking stopped being a consumer points product. It is now closer to infrastructure procurement. Teams pay for economic security. You decide whether the premium covers the extra failure paths.

That shift is why headlines look worse than the core design. TVL fell after slashing became real. The remaining capital is more honest about risk.

How restaking went from a giant narrative to a smaller working market

EigenLayer launched on Ethereum in 2023. It let validators and LST holders opt into extra work.

2024 turned that into a retail product. Liquid restaking tokens made the position tradable. Points programs pulled in deposits. Sector TVL printed huge dollar figures during that window.

Two dates changed the tone.

On 17 April 2025, EigenLayer activated production slashing with Unique Stake and Operator Sets. An AVS could finally penalize a defined slice of an operator’s allocation.

On 17 June 2025, Eigen Labs introduced EigenCloud. Restaking stayed the security base. The new pitch was a verifiable cloud with EigenDA, EigenVerify, and EigenCompute. a16z crypto bought an additional $70 million of EIGEN around that launch.

Other venues arrived in the same cycle. Symbiotic launched a permissionless vault model with slashing from day one. Babylon launched Bitcoin native staking and the BABY token. Karak later left the race as a restaking peer and rebranded toward OpenGDP.

April 2026 was the stress test. Kelp DAO’s rsETH stack lost about $292 million through a cross chain verification failure. Redemptions then hit the wider restaking market.

August 2026 closed a chapter for retail wrappers. ether.fi removed restaking from weETH and moved restaking exposure into weETHs on Symbiotic. Less than 1% of ether.fi assets still sat on EigenLayer at the time of that disclosure. The plan was to reach zero in the third quarter of 2026 and to drop EigenPod withdrawal credentials in the fourth. Late August is still inside that window. Treat zero as a target until the issuer page says it is done.

On 26 August 2026, Polychain moved 14.65 million EIGEN, about $3.09 million, to Coinbase Prime. That flow came from a larger May redemption. It is a reminder that token supply and protocol TVL do not move together.

How restaking works under the hood

You meet four roles.

• Staker: you supply ETH, an LST, BTC, or another accepted asset.
• Operator: a node business that runs the extra software.
• Service: the AVS, network, or Bitcoin secured chain that wants protection.
• Marketplace: the contracts that track allocations, rewards, and slashing.

An optional fifth role is the liquid issuer. That issuer takes your deposit, restakes it, and gives you a receipt token.

Native restaking versus LST restaking

Native restaking uses withdrawal credentials. On EigenLayer that path runs through EigenPods. You keep a validator shaped position and opt the same stake into extra services.

LST restaking deposits a liquid staking token such as stETH, rETH, or cbETH. You avoid running a validator. You add the LST issuer’s contracts to your risk stack.

Neither path is free of operational risk. Native restaking concentrates key and operator risk. LST restaking adds another protocol and a second withdrawal clock.

Unique Stake, Operator Sets, and redistributable slashing

Older explainers warned that one AVS fault could slash your entire restaked pile. The 2025 design tried to stop that cascade.

An AVS now defines Operator Sets. An operator joins a set and accepts that set’s rules. Unique Stake earmarks a slice of collateral to that set. Another AVS cannot seize that slice.

Slashing is still opt in at three levels.

1. The service turns slashing on.
2. The operator registers for that slashable set.
3. You, or your liquid issuer, choose that operator.

Later upgrades added more process. Redistributable slashing can send penalized funds to a recipient instead of only burning them. A resolution delay of about seven days sits in front of final clearing.

Ask a sharper question than “is slashing live?” Ask what share of your dollars sits in a slashable set, for which offenses, with what delay.

ether.fi’s public slashing page said it did not sign up to slashable services. It also said the remaining EigenLayer restake share was under 1% in August 2026, with a plan to reach zero in the third quarter.

Exits, queues, and liquidity buffers

Your exit path depends on the wrapper.

Native EigenLayer withdrawals face protocol delays after you undelegate. Ethereum validator exits face a separate consensus queue. LST and LRT issuers can pay some redemptions from a liquid buffer. That buffer is finite.

After the April 2026 rsETH shock, ether.fi said it processed 542,792 ETH of withdrawals, about 19.6% of TVL, in 33 days. It used consolidation operations rather than dumping every validator into Ethereum’s shared exit queue. Median claim time was about 4.9 days.

That episode matters. Liquidity engineering can protect you in a run. It cannot repair stolen backing.

EigenLayer versus EigenCloud

People still use the names as synonyms. They are not.

EigenLayer is the restaking marketplace. It holds collateral, operators, allocations, rewards, and slashing.

EigenCloud is the product layer announced in June 2025. It packages first party services on top of that marketplace.

The first party set includes:

• EigenDA for data availability.
• EigenVerify for dispute resolution, including objective and intersubjective modes.
• EigenCompute for verifiable offchain containers.

The public pitch later widened toward verifiable apps and agents. Coinbase’s asset page now describes EigenCloud as a developer platform for trust and verifiability on or off chain.

This split changes how you judge the system. Restaking yield may eventually come from application fees, not from endless points. That fee stream is still the open question. TVL alone does not answer it.

Eigen Labs also cut about 25% of staff in July 2025 to focus on EigenCloud. That is useful context. The company chose a narrower product bet after the restaking boom.

Restaking beyond Ethereum

EigenLayer still leads Ethereum restaking by secured value. It is no longer the only serious design.

July 2026 comparison pieces put EigenLayer near $5.0 billion of base layer TVL, Babylon near $3.3 billion in BTC terms, and Symbiotic near $329 million on DefiLlama’s base layer print. Those prints exclude some liquid wrappers and move with asset prices.

Those July figures come from comparison writeups that cited DefiLlama base-layer prints. They exclude some liquid wrappers. They move with asset prices.

Your collateral usually chooses the venue. Solayer sits in the Solana bucket with Jito. It is not an EigenLayer fork. ETH goes to EigenLayer or Symbiotic. Native BTC goes to Babylon. SOL goes to Jito or other Solana systems.

Symbiotic

Symbiotic uses permissionless ERC-4626 vaults. Networks set their own slashing and rewards. Collateral is not limited to ETH and LSTs.

That flexibility is the pitch. The tradeoff is a smaller operator and service graph than EigenLayer. There is also no widely traded protocol token yet, so incentives still lean on points and partner rewards.

ether.fi’s 2026 restaking sleeve moving toward Symbiotic is the adoption signal to watch. A large issuer changing venues is more informative than a new logo.

Babylon

Babylon locks native bitcoin in self custodial scripts. You do not wrap BTC on Ethereum to start. Finality providers use that locked BTC as economic weight for other systems.

This is the cleanest “beyond EigenLayer” story. The risk set is different. You care about Bitcoin script safety, unbonding time, and finality provider quality. You do not inherit EigenLayer Operator Sets.

TVL here swings with the bitcoin price. Compare Babylon in BTC terms as well as dollars.

Jito and Solana restaking

Jito restaking mints vault tokens against staked Solana assets. Those vaults secure Node Consensus Networks, Solana’s analogue to AVSs. TipRouter was an early network used to decentralize MEV tip distribution.

The restaking sleeve itself stayed small through mid 2026, often cited in the low tens of millions of dollars. JitoSOL and Jito’s broader Solana stack are much larger. Do not confuse liquid staking scale with restaking scale.

Karak as a post mortem

Karak launched as a multi asset, multi chain restaking venue. By July 2026, restaking deposits tied to the old brand sat near $6.5 million after a pivot to OpenGDP.

Leave it in the history section. Do not treat it as a current top three option.

Liquid restaking after the unbundle

A liquid restaking token is a receipt. You deposit ETH or an LST. The issuer restakes that capital. You receive a tradable token.

In 2024 that receipt also meant “this token is the EigenLayer trade.” That sentence broke in 2026.

ether.fi made weETH a plain liquid staking token again. 

Restaking moved to weETHs on Symbiotic. CEO commentary called the bundled era over.

The demand print mattered more than the press line. After the split, weETHs was reported around 9,136 tokens, near $18 million, about 0.5% of ether.fi’s staking book. When restaking stopped hiding inside weETH, almost all of the money stayed in plain staking. That is the 2026 signal. TVL from 2024 is not.

Read ticker pages slowly after that change. Some older docs still describe eETH or weETH as natively restaked. The slashing risk page and August 2026 product split are the current source of truth.

Other issuers still package restaking inside one token. Renzo’s ezETH, Kelp’s rsETH, and Puffer’s pufETH remain part of that market. Compare each issuer’s operator policy, slashable set exposure, and bridge design. Do not assume they share ether.fi’s new split.

Two kinds of depeg

A liquidity depeg is a market discount. The backing can still exist. Sellers overwhelm thin pools. ezETH’s 2024 air drop shock is the textbook case.

A backing loss depeg is different. Assets leave the system. Every token is worth less even if you wait. The April 2026 rsETH event sits in this bucket.

If you use an LRT as lending collateral, both depegs can liquidate you. Backing loss is worse. The discount may not mean revert.

The Kelp case study

On 18 April 2026, attackers abused a verification weakness in Kelp’s cross chain setup. Reporting put the drain near $292 million. About 116,500 rsETH left the normal backing path and was reused as collateral, including on Aave. Researchers tied the theft to North Korea’s Lazarus Group.

The path was a single-verifier, 1-of-1 DVN route on the cross-chain hop. That is a wrapper design choice. It is not an AVS slashing rule.

This was not an EigenLayer slashing event. The break sat in bridge verification and then jumped into lending markets. rsETH markets froze on venues such as Aave and SparkLend.

Earlier Aave governance had added a high LTV e-mode path for rsETH. Incident writeups later put possible Aave-market bad debt in a wide band, about $124 million to $230 million, depending on how losses were socialized. Measure a position by hops to a lending market, not by advertised APY. 

That setting helped the token circulate as collateral. It also transmitted the shock.

If you want one 2026 lesson, use this one. Restaking risk showed up as rehypothecation, not as a textbook AVS slash.

What you actually earn in 2026

Early restaking ads stacked consensus yield, points, and partner emissions into one large APY. Separate those layers now.

A conservative ETH stack is base staking plus a large LST. A core restaking stack adds a disclosed extra coupon. An aggressive stack loops the receipt token.

Loop APYs can look like 15% or more. Those numbers require a stable peg and a stable borrow rate. They failed that test in April 2026 for rsETH users.

Bitcoin and Solana coupons follow their own issuance and fee markets. Babylon rewards mix BTC lock economics with BABY incentives. JitoSOL yield includes Solana inflation and MEV tips. Neither should be pasted into an EigenLayer APY table.

Sector fee prints from early 2026 showed restaking producing measurable revenue, including estimates above $527,000 a day across the category in late February. That is useful. It is still small next to peak TVL.

The restaking risk stack

Score a position before you deposit. A high coupon that fails several of these tests is not a bargain.

1. Base-chain staking slashing: Ethereum, Bitcoin, or Solana already has penalty rules.
2. Service slashing: Extra offenses live in AVS or NCN rules.
3. Operator quality: Keys, uptime, and client diversity sit with someone else if you delegate.
4. Marketplace contracts: Bugs here hit every restaker on that venue.
5. Wrapper accounting: Share price, oracles, and vault logic can leak value.
6. Bridges and messaging: Kelp showed a single verifier design can mint a sector crisis.
7. Lending liquidation: Aave-style e-mode turns a depeg into forced selling.
8. Exit time: Queues grow when everyone leaves.
9. Upgrade keys and governance: A small council can change risk after you deposit.
10. Concentration. One marketplace, a few operators, and one collateral listing on a big lender.
11. Incentive withdrawal: When points end, mercenary deposits leave.
12. Custody and policy: Institutions need withdrawal control and written slash limits.

Give yourself a slash budget. Decide the largest percentage of net worth you can lose to a service you do not control. If the product cannot show slashable set exposure on chain, size it as if the budget is already in use.

On 26 August 2026, Ethereum researchers also floated a first step toward quantum-resistant validator deposits. That work targets the deposit contract and today’s BLS-heavy flow. Restaking inherits Ethereum’s long-range cryptography assumptions. It is a horizon issue, not a next-week trade.

How to restake in 2026

Do not start with a ticker. Start with an exit plan and a slash budget.

Path A: native EigenLayer

Use this path if you already run validators or can evaluate operators.

1. Confirm withdrawal credentials and EigenPod setup in official docs.
2. Choose operators with public slashing pages and clear Operator Sets.
3. Read whether those sets are slashable, redistributable, or both.
4. Allocate only a slice you can leave idle through undelegation delays.
5. Track rewards per set, not as one blended APY.

Native restaking gives you more visibility. It also makes operator failure your problem more directly.

Path B: a liquid token

Use this path if you want a receipt you can sell or post as collateral.

1. Identify whether the token is an LST, an LRT, or a split pair such as weETH and weETHs.
2. Read the issuer’s current slashing policy, not a 2024 blog post.
3. Check bridge routes. Avoid designs that rely on a single verifier.
4. Check lending listings and LTVs if you might borrow against the token.
5. Size the position for a multi day exit, not a same hour exit.

If the issuer hides restaking inside a familiar LST name, assume most holders do not understand the risk.

A simple sizing rule: if you cannot explain the slashable set in one sentence, cap the wrapper at a slice you can leave idle through a multi-day exit. For most readers that is a satellite position, not the core ETH stack.

Path C: Babylon and native bitcoin

Use this path if you want BTC productive without wrapping it first.

1. Use official Babylon flow for the timelock script.
2. Select finality providers the way you would select an operator.
3. Record the unbonding window, often described around 1,008 Bitcoin blocks.
4. Separate BABY incentives from BTC principal risk.
5. Do not compare the coupon one for one with ETH restaking APY.

Bridged or wrapped bitcoin products are a different tool. They reintroduce the wrap and bridge risks Babylon tries to avoid.

Pre deposit checklist

• Can you see slashable sets on chain?
• Is redistribution enabled for those sets?
• Who holds signing keys, and do they use DVT or hardware isolation?
• How long is the panic exit?
• Is the token listed as high LTV collateral?

If you cannot answer those questions, you are not restaking. You are buying a yield ticker.

How many hops sit between this token and a high LTV lending market?

Who should restake, and who should not

Institutional flow is the quiet part of 2026. Public examples include SharpLink Gaming routing treasury ETH through EigenCloud and ether.fi rails, Flow Traders acting as an operator on an EigenLayer credit market, and Hex Trust routing bitcoin into Babylon. Those desks buy process. They do not buy points.

Most readers should keep restaking as a satellite position. The base stake still does the economic work that matters.

EIGEN, other tokens, and value capture

You do not need EIGEN to restake ETH. That single fact explains a lot of the token’s 2026 tape.

EIGEN traded near $0.1975 on 26 August 2026, with a market cap near $173 million and an all time high of $5.65. Two days later, on 28 August, it still sat near $0.19 to $0.20, with a market cap near $170 million to $175 million. One year earlier it sat near $1.23 to $1.26, depending on the feed.

EigenCloud can process fees while the token still struggles to capture them. Commentaries through July 2026 kept repeating that split. Protocol TVL in the several billion range sat beside a token down about 95% from its high.

ELIP-12 is the live attempt to close the gap. The proposal routes a 20% fee on subsidized AVS rewards and 100% of EigenCloud service fees toward possible EIGEN buybacks. Passage and actual buyback flow are separate events. Unlocks also continue.

The 26 August Polychain transfer to Coinbase Prime is a current supply exhibit, not a morality play. Of 131.8 million EIGEN redeemed at the end of May, 46.86 million were restaked, 14.65 million moved that day, and 70.29 million remained in related addresses.

Babylon’s BABY token and Symbiotic’s no token stance are different bets. Emissions can bootstrap usage. They can also mask a thin fee market. Judge each token by whether service fees need the token at all.

What restaked capital secures now

Ignore vanity counts such as “190 AVSs in development.” Count live jobs that pay or at least slash. If the coupon is still emissions, the insurance premium is being paid by a token printer.

Common jobs in 2026:

• Data availability for rollups, led on this stack by EigenDA.
• Dispute resolution and verification markets.
• Offchain compute with attestations.
• Oracles and interoperability networks.
• Solana MEV and tip distribution through NCN style networks.
• Bitcoin anchored finality for other chains.

The honest metric is fee paying demand. If services cannot pay restakers from revenue, the coupon falls back to emissions. Emissions end. Revenue either shows up or it does not.

Tax, custody, and institutional reality

This is education, not tax or legal advice.

Extra rewards are often income when you receive them, depending on your jurisdiction. Selling an LRT during a depeg can also create a taxable disposal. A slash may not look like a simple capital loss. Talk to a professional before you treat any of those events as routine staking.

Custody questions are sharper than yield questions.

• Who controls withdrawal credentials?
• Can your custodian refuse slashable Operator Sets?
• Are restaking rewards swept to a segregated account?
• What report will you receive after a slash or a wrapper exploit?

If those answers are fuzzy, use plain staking. Shared security is a poor place to hide operational debt.

What to watch next

• EigenCloud fees and any ELIP-12 buyback flow.
• Whether other LRT issuers copy ether.fi and split LST tickers from restaking tickers.
• Aave and other lenders tightening e-mode after the rsETH shock.
• The first material EigenLayer slash, including burn versus redistribution and LRT NAV math.
• Babylon’s BTC locked amount versus wrapped bitcoin products.
• Solana NCNs finding demand beyond tip routing.
• Operator tooling such as anti slashers and state aware signing guards.
• Ethereum’s quantum resistant deposit research, which sits on a multi year horizon.

Sources and references

Last updated 28 August 2026. Dates in the article are snapshots. If a protocol page and an aggregator disagree, the protocol page wins for policy.

Protocol pages

• ether.fi, “Slashing Risk,” August 2026. Under 1% still on EigenLayer; no slashable AVS opt-in; Q3/Q4 wind-down.
• ether.fi, “weETH split announcement,” 6 August 2026. weETH as LST; restaking moved to weETHs on Symbiotic.
• ether.fi, “April 2026 withdrawal note.” 542,792 ETH processed after the rsETH shock.
• EigenCloud, “Slashing concept.” Unique Stake, Operator Sets, redistributable slashing.
• Eigen Foundation, “ELIP-002: Slashing.”
• EigenLayer Forum, “Introducing EigenCloud,” 17 June 2025. Includes the a16z crypto $70 million EIGEN purchase.
• EigenCloud, “ELIP-12 value-capture note,” January 2026. 20% subsidized reward fee and 100% EigenCloud fees toward possible buybacks.

Market snapshots

• DefiLlama, “EigenLayer.” Refresh on publish day.
• Protofire, “EigenLayer vs Symbiotic vs Babylon: Restaking Compared (2026),” July 2026. Source of the ~$5.0B / ~$3.3B / ~$329M comparison print.
• MetaMask, “EIGEN price.” 26–28 August 2026 tape.
• Coinbase, “EIGEN / EigenCloud asset page.”

Reporting

• The Defiant, “ether.fi Removes Restaking From weETH,” 6 August 2026.
• The Block, “Kelp DAO’s rsETH bridge exploited for roughly $292 million,” 18 April 2026.
• Crypto Briefing, “Aave implications of the Kelp exploit.” Bad-debt range cited in the article.
• Odaily, “Polychain transfers 14.65 million EIGEN to Coinbase Prime,” 26 August 2026.
• The Block, “SharpLink stakes ETH on Linea with EigenCloud and ether.fi,” January 2026.
• CoinDesk, “Eigen Labs axes 25% of staff to focus on EigenCloud,” 9 July 2025.
• CoinDesk, “Ethereum developers propose first step to protect ETH staking from quantum attacks,” 26 August 2026.

Disclaimer

This article is only for educational purposes. It is not tax, legal, or investment advice. Restaking can lose principal. Slashing, wrapper bugs, bridges, and lending liquidations are all live paths. Dated TVL and price lines are snapshots, not forecasts. Nothing here is an offer to buy or sell EIGEN, weETH, weETHs, rsETH, BTC, or any other asset. Named protocols, operators, and institutions are examples, not recommendations. If you cannot name the slashable set, the exit clock, and the largest loss you will accept, stay in plain staking.