What is restaking? Liquid restaking, EigenLayer, and the new yield stack explained

Restaking lets you take the ether you have already staked and put it to work a second time, securing other protocols for extra yield. It is one of the largest ideas in crypto today, and one of the most misunderstood. Here is how it works, where the yield comes from, and what can go wrong.

Restaking is the practice of taking crypto that has already been staked to secure a blockchain, usually ether on Ethereum, and committing that same stake a second time to help secure additional protocols, in exchange for extra yield and extra risk.

That is the core idea in one sentence, and it is both simpler and stranger than it sounds. Normal staking locks up your ether to help secure Ethereum and pays you a reward for doing so. Restaking reuses that same staked capital, pointing it at other services that also need security, so a single pool of ether can earn from more than one source at once. The concept was introduced by a protocol called EigenLayer in 2023, and by 2026, it had grown into one of the largest systems in all of decentralized finance, with tens of billions of dollars committed to it.

This guide explains restaking from the ground up, with no assumption that you already understand staking deeply. It covers what restaking actually is and the problem it solves, how it works under the hood through restakers, operators, and the services they secure, the liquid restaking tokens that most people actually use, a worked example of how the yield stacks up, the real risks including a major 2026 exploit that reset the conversation, the competitors challenging EigenLayer, and why restaking has come to be seen as infrastructure, not a passing yield craze. By the end, you will understand not just the mechanics but the trade-off at the heart of it: more yield, layered on top of more ways to lose money.

What restaking actually is

To understand restaking, you first need a clear picture of ordinary staking. Ethereum is secured by validators, participants who lock up ether as collateral and are rewarded for honestly processing transactions. If a validator misbehaves, part of its stake can be taken away, a penalty called slashing. This staked ether is the economic backbone of the network: attacking Ethereum would mean risking enormous sums, which is what keeps it safe. Base staking pays a reward of roughly three to four percent a year for providing that security.

Restaking asks a clever question: that staked ether is sitting there securing Ethereum, so why can it not also secure other things at the same time? Many newer protocols, such as services that help blockchains store data, bridges that move assets between networks, and oracles that feed outside information on-chain, need their own economic security. Building that security from scratch, by launching a token and recruiting a validator set, is slow and expensive. Restaking lets these services rent security instead, borrowing the trust of Ethereum’s already-staked capital. A useful way to picture it: if your shop sits inside a mall that already has guards at every entrance, you do not need to hire your own. But a shop in a standalone building has to arrange its own security. Restaking is the mall guards agreeing to watch the standalone shops too, for extra pay, while accepting that they can be penalized if they fail at either job.

From staking to restaking: the problem it solves

The deeper reason restaking exists is a problem in how new blockchain services bootstrap trust. Any protocol that needs participants to behave honestly faces a chicken-and-egg dilemma: it needs a large amount of valuable, at-risk capital to make attacks costly, but attracting that capital requires already being valuable and trusted. A brand-new data-availability layer or oracle network starts with neither. Historically, each such service had to launch its own token, convince people to stake it, and hope the token held enough value to deter attackers, a fragile and capital-hungry process that often failed.

Restaking collapses that problem. Instead of every new service building its own security pool, they all draw on a shared one made of restaked ether, the most liquid and trusted crypto collateral after Bitcoin. The services that consume this shared security are called Actively Validated Services, usually shortened to AVSs. An AVS specifies what it needs validated and what rules its participants must follow, and in return, it pays rewards to the restakers who back it. For the broader ecosystem, the appeal is obvious: new protocols launch with strong security on day one, and stakers earn more on capital that was already committed. This is why restaking is often described as a marketplace for trust, matching services that need security with capital willing to provide it for a fee.

How restaking works under the hood

Three roles make the system run, and keeping them straight removes most of the confusion. First are restakers, the people who supply the capital by depositing ether or staked-ether tokens into the restaking protocol and opting in to its additional rules. Second are operators, the entities that actually run the technical software each AVS requires, validating data, sequencing transactions, or whatever the service needs. Most restakers do not run anything themselves; they delegate their stake to an operator they trust, much as a token holder might delegate a vote. Operators typically take a cut, around ten percent, of the rewards they earn for the restakers who back them. Third are the AVSs themselves, the services being secured.

The mechanism that ties it together is slashing, the same penalty that secures Ethereum, extended to cover the extra commitments. When a restaker opts in to back an AVS through an operator, that stake becomes subject to the AVS’s slashing conditions on top of Ethereum’s own. Behave honestly, and you earn rewards from both Ethereum and the AVS. Misbehave, or delegate to an operator who does, and you can be slashed on both. By 2026, slashing on EigenLayer is live on the main network, and the protocol has even added the ability for slashed funds to be redistributed rather than simply destroyed, which raises the stakes further.

Withdrawals are not instant either; exiting a restaking position involves a delay of several days, a safety feature that also means you cannot flee instantly in a panic. The whole arrangement is a layered bet: each AVS you back adds a reward stream and a matching way to lose your collateral.

Liquid restaking and the tokens most people use

For most users, the description above is more complex than they want to manage, which is where liquid restaking comes in, and it is how the large majority actually participate. Rather than depositing ether directly into EigenLayer, choosing operators, and tracking AVS allocations yourself, you deposit into a liquid restaking protocol that handles all of it. In return, you receive a liquid restaking token, or LRT, a single token that represents your restaked position and quietly accrues all the layered yield. The leading protocols are ether.fi, which issues a token called eETH, along with Renzo with ezETH, Kelp DAO with rsETH, and Puffer with pufETH. A handful of these protocols hold the majority of all restaked ether.

The word “liquid” is the key benefit. A normal restaking position is locked and illiquid, but an LRT is a freely tradeable token, so you can hold it, sell it, or use it elsewhere in decentralized finance while it keeps earning. Think of an LRT as a receipt that itself earns yield: you hand over your ether, the protocol restakes it across services on your behalf, and the receipt you get back both represents your claim and grows in value as rewards accrue. This composability, the ability to use the receipt as collateral or in liquidity pools, is what turned restaking from a niche activity into a core building block of DeFi. It is also, as the next sections show, where some of the sharpest risks hide, because a receipt is only as sound as the system standing behind it.

A worked example: stacking the yield

Make it concrete. Suppose you hold ten ether and want to put it to work. You deposit it into a liquid restaking protocol like ether.fi and receive ten units of its LRT in return. From that moment, three layers of yield begin stacking. The first layer is base Ethereum staking, paying roughly three to four percent a year, because your ether is still securing Ethereum underneath everything. The second layer is the AVS rewards, the extra payment from the services your stake now also secures, typically adding one to two percent. The third layer, common in 2026, is points or token incentives, where protocols hand out their own tokens to attract deposits. Add the real yield layers together, and you land somewhere around four to seven percent, meaningfully above plain staking.

Some users push further with a technique called looping, and it is essential to understand why it is dangerous. Because the LRT is a liquid token, you can deposit it as collateral on a lending protocol, borrow more ether against it, buy more LRT with that borrowed ether, and repeat. Each loop multiplies your exposure to the yield, and headline returns of twelve to twenty percent become possible. But each loop also multiplies your exposure to loss, because you now carry a loan that can be liquidated if prices move against you or if the LRT temporarily trades below the value of the ether it represents. Looped restaking is not passive income; it is leveraged yield farming with extra steps, and the same mechanism that magnifies the gain magnifies the wipeout. One more caveat matters: a large share of restaking yield in 2026 still comes from protocols handing out their own tokens, not from real fees paid by AVSs, which means some of that attractive yield is inflationary and may compress as emissions slow.

The risks: slashing, smart contracts, and concentration

Restaking adds yield by adding risk, and the risks are not theoretical. The most obvious is slashing: by opting in to secure additional services, you accept additional ways to be penalized, and a poorly run operator can lose your funds through simple negligence instead of malice. The second is smart-contract risk, and restaking stacks it in layers. Your capital passes through Ethereum’s staking contracts, then EigenLayer’s contracts, then often a liquid restaking protocol’s contracts, and a bug anywhere in that chain can drain funds. More layers mean more places for something to break. A third risk is liquidity: during market stress, an LRT can trade below the value of the ether backing it, so trying to exit in a panic can mean selling at a discount.

The most consequential risk is concentration, and 2026 delivered a hard lesson in it. EigenLayer holds roughly ninety-four percent of the restaking market, which means a serious exploit there would not be one protocol’s problem; it would be a systemic event for everything secured by restaked ether, including the data layers that major rollups depend on. The fragility showed up in April 2026, when Kelp DAO, one of the large liquid restaking protocols, suffered an exploit of around three hundred million dollars that triggered roughly five and a half billion dollars in withdrawals across the entire restaking sector as nervous depositors fled. The protocol survived, but the episode reset how seriously the industry takes smart-contract risk in restaking. The lesson is not that restaking is doomed, but that the extra yield is genuine compensation for genuine danger, and treating it as free money is the surest way to get hurt.

Beyond Ethereum: the competition and the bigger vision

EigenLayer started the category, but by 2026, it was no longer alone, and the alternatives reveal where restaking is heading. Symbiotic positions itself as a fully permissionless, modular competitor that accepts not just ether but a wide range of tokens as restakable collateral, an approach called asset-agnostic restaking. Karak pushes asset diversity even further, supporting stablecoins, wrapped Bitcoin, and liquidity-provider tokens. Separately, a protocol called Babylon brought the restaking idea to Bitcoin, letting Bitcoin holders provide security to other networks. The competition is increasingly about which assets you can restake and how flexibly, not just about ether on Ethereum.

EigenLayer itself has expanded its ambitions well beyond yield, and this is the part that hints at restaking’s longer significance. Its flagship service, EigenDA, is a data-availability layer that helps rollups store transaction data cheaply, and it remains the single largest consumer of restaked security. The protocol has since added services for verifiable AI inference and off-chain compute, rebranding the whole effort as a kind of decentralized cloud built on rented trust. Whether that vision holds is an open question, because much of the capital flowed in chasing yield, not because the underlying services already generate large fees. The bull case is that restaking becomes permanent infrastructure, the security layer beneath a new generation of decentralized services. The bear case is that the fees never grow into the valuations, and the yield that attracted the capital fades. Both remain possible.

What to check before you restake

Because restaking layers risk on risk, a short checklist separates a reasonable position from a reckless one, and running through it before committing funds is worth the few minutes it takes. Start with the protocol itself: favor names that have been audited by reputable firms, ideally more than once, and that have operated through real market stress without losing user funds. A protocol’s age and track record are not guarantees, but a system that has survived volatility and scrutiny is safer than a new one promising the highest yield. Be especially wary when an unfamiliar protocol advertises returns far above the established players, because outsized yield is usually paying you for outsized risk you cannot see.

Next, understand the lockup before you deposit, not after. Restaking withdrawals is not instant; exiting can take days, and during a market panic, that delay can trap you in a position you want to leave. Know the exit timeline and make sure you will not need the funds inside it. Then look at where your yield actually comes from: if most of it is the protocol handing out its own token, treat that as inflationary and temporary, separate in your mind from the steadier yield paid by real service fees. A headline number propped up by token emissions can shrink the moment those emissions slow. Check, too, whether the protocol keeps any kind of insurance fund or buffer against slashing, since some set aside a small share of rewards for exactly that purpose.

Finally, size the position sensibly and diversify. Gas fees on Ethereum make tiny restaking positions uneconomic, so the rewards need to justify the transaction costs, which argues for a meaningful but not reckless amount. Spreading across more than one protocol reduces the damage if any single one is exploited, a lesson the 2026 sector-wide withdrawal scare drove home. Above all, treat any liquid restaking token as only as sound as the layers beneath it, and never deposit money you could not afford to see slashed, stuck through a withdrawal delay, or caught in a depegged token during a sell-off. The providers who do well in restaking are the ones who treat the checklist as mandatory, not optional.

Why restaking matters

Strip away the jargon, and restaking is a single, powerful idea: that valuable, at-risk capital securing one network can be reused to secure many, turning Ethereum’s staked ether into shared security that any new protocol can rent. That idea changed how blockchain services launch, letting them inherit strong security on day one instead of building it painfully from scratch, and it gave ether holders a new way to earn on capital that was already working. In a few years, it grew from a concept into one of the largest systems in crypto, which is why it is now treated as infrastructure rather than a fad.

The honest conclusion is that restaking is real and significant, and also genuinely risky in ways that are easy to underestimate. The yield is not magic; it is payment for taking on layered slashing, stacked smart-contract exposure, and the systemic fragility that comes from so much value sitting in so few protocols, as the events of 2026 made plain. For anyone considering it, the framework is simple to state and harder to live by: understand every layer your capital passes through, treat extra yield as a signal of extra risk, favor audited and battle-tested protocols, and never commit funds you cannot afford to see slashed or stuck. Restaking rewards those who respect what it actually is, a sophisticated trade-off, and punishes those who mistake it for a free lunch.

Frequently Asked Questions

This guide is educational information, not financial advice. Restaking involves slashing, smart-contract, and liquidity risks that can lead to loss of funds, and figures and protocol details reflect the state of the market as of June 24, 2026. Use audited, well-tested protocols and never commit more than you can afford to lose.