Restaking Explained: Yield, Risk, and Ethereum’s New Trust Layer
Restaking lets staked ether secure extra protocols for extra yield, and it now anchors billions of dollars in DeFi. Here is how it works, what it pays, and where the risks really hide.
Two years ago, restaking was a niche pitch at Ethereum side events. Today it anchors one of the largest and most argued-over corners of decentralized finance, with billions of dollars of staked ether pulling double duty as security for other protocols. The sales line is short enough to fit on a sticker: stake once, earn more than once. The reality is a stack of moving parts, a fresh set of risks, and a regulatory picture that is still only half drawn. Here is what restaking is, how the money actually flows, and why even Ethereum’s co-founder has urged people to slow down.
What restaking actually is
Restaking means taking ether that is already staked, or a liquid staking token that represents staked ether, and committing it a second time as collateral to secure additional services. The idea was popularized by EigenLayer, which built a marketplace where stakers lend their economic weight to other protocols in exchange for extra rewards. CoinDesk’s primer puts it plainly: restaking uses already-staked crypto to help secure networks and services beyond Ethereum itself (CoinDesk).
The appeal is capital efficiency. Ethereum’s base layer is guarded by tens of billions of dollars of staked ether that, for the most part, does nothing beyond validating blocks. Restaking asks an obvious question: why can that same collateral not also backstop an oracle, a bridge, or a data availability layer, and collect fees for the trouble? If the answer is that it can, then one pool of capital secures many things at once. That single insight turned EigenLayer, which opened to depositors in 2023, into one of the fastest-growing applications DeFi had seen.
How it works: stakers, operators, and AVSs
Restaking runs on three roles. Restakers supply the capital. Operators run the node software that does the actual work. Actively Validated Services, usually shortened to AVSs, are the protocols buying security. The flow is straightforward once it is laid out.
- Restakers deposit ether or a liquid staking token and delegate it to an operator.
- Operators validate for one or more AVSs and take a cut of the rewards.
- AVSs (think data availability layers, oracles, bridges, and co-processors) rent that pooled security instead of bootstrapping their own validator set.
Because the same restaked ether can back several AVSs at once, EigenLayer calls the model pooled security. Behave honestly across all of them and you earn from Ethereum and from each service. Misbehave, or delegate to an operator who does, and you can be penalized on every front at the same time. That two-way street is the whole point, and also the whole problem.
Liquid restaking tokens and the new yield stack
Most people never touch the raw machinery. Instead they use liquid restaking platforms such as ether.fi, Renzo, Kelp DAO, and Puffer, which accept deposits, restake on the user’s behalf, and hand back a tradable receipt called a liquid restaking token, or LRT. Examples include weETH from ether.fi, ezETH from Renzo, and rsETH from Kelp DAO. These tokens keep accruing rewards while staying liquid, so holders can lend them, post them as collateral, or farm them elsewhere. Renzo, Kelp DAO, and Puffer compete on which assets and chains they support and on fees, and each issues its own branded receipt.
That is the yield stack in action: base staking rewards, plus restaking rewards, plus whatever a holder can earn by putting the LRT to work across DeFi, often topped with loyalty points that may convert to future tokens. ether.fi has led the category and markets a feature that lets users keep control of their own validator keys rather than handing them over. The convenience is real, but every layer is one more smart contract that has to hold, and one more peg that has to be defended.
Where restaking stands in 2026
At its 2024 peak, restaking deposits approached $20 billion and briefly pushed EigenLayer past lending heavyweight Aave in total value locked (Blockworks). The sector has cooled and matured since. Trackers put restaking deposits somewhere between the high single digits and the mid-teens of billions of dollars in 2026, with the spread depending on how you count; DeFiLlama, for example, records the underlying stake once at the base layer to avoid counting the same ether twice. Much of the early surge was driven by users farming anticipated token airdrops; as those incentives cooled, deposits drifted lower and the capital that stayed looked stickier.
On the consumer side, a handful of liquid restaking tokens hold most of the value. The figures below are approximate market capitalizations in US dollars as of mid-2026.
| Token | Issuer | Approx. market cap (USD) | Notes |
|---|---|---|---|
| weETH | ether.fi | ~$2.8 billion | Largest LRT by value |
| rsETH | Kelp DAO | ~$870 million | Target of the April 2026 bridge exploit |
| LBTC | Lombard | ~$615 million | Bitcoin-based restaked collateral |
| eETH | ether.fi | ~$200 million | Unwrapped ether.fi receipt token |
| ezETH | Renzo | ~$80 million | Multi-chain restaked ether |
Slashing: the risk that stacks
Slashing is the penalty a network imposes when a validator or operator breaks the rules, and it is the mechanism that makes restaking economically meaningful. The trade is blunt: by opting in to back an AVS, a restaker’s collateral becomes subject to that service’s slashing conditions on top of Ethereum’s own. EigenLayer switched on slashing in 2025, finally completing the design it launched with (CoinDesk). More yield, in other words, arrives with more distinct ways to lose principal.
The risks do not just add up; they correlate. A bug in an LRT contract, an operator that grows too large, or a single faulty AVS can ripple through pools that all lean on the same collateral. That is a different animal from plain staking, where the main thing you worry about is your own validator’s uptime.
The Kelp DAO exploit and the off-chain attack surface
In April 2026 the abstract risk turned concrete. An attacker minted roughly 116,500 rsETH with nothing backing it, draining about $292 million from Kelp DAO and producing the largest crypto exploit of the year (CoinDesk). Crucially, the flaw lived in off-chain bridge infrastructure, a single-verifier cross-chain setup, rather than in the core restaking contracts. The attack surface, it turns out, reaches well beyond Solidity.
The cleanup was messy. Stolen ether scattered across roughly 20 chains, and Arbitrum moved to freeze about $71 million tied to the theft (CoinDesk). The episode dented confidence across the sector and drove home an uncomfortable lesson: a liquid restaking token inherits the weaknesses of every bridge, oracle, and node it quietly depends on. In the days that followed, nervous holders pulled funds from rival platforms as well, a reminder that fear in this corner of DeFi travels fast.
Vitalik’s warning about overloading Ethereum
Not everyone is comfortable with how far this goes. As early as 2023, Ethereum co-founder Vitalik Buterin published an essay titled Don’t overload Ethereum’s consensus that drew a careful line (vitalik.eth.limo). Reusing staked ether for extra duties is broadly fine, he argued; trying to recruit Ethereum’s social consensus to rescue an outside application is not.
The worry is systemic. If enough money rides on restaking and a major service fails or is attacked, there could be pressure for a contentious hard fork to make victims whole, dragging Ethereum’s base layer into fights that were never its own. Restaking advocates answer that clear contracts and bounded slashing keep the damage with the people who opted in. Which camp is right may stay unsettled until the system is genuinely stress-tested at scale.
The competition: Symbiotic, Karak, and multi-asset restaking
EigenLayer started the category but no longer owns it outright. Symbiotic, an early Paradigm-backed project, went live on Ethereum mainnet in early 2025 as a permissionless, modular restaking protocol that accepts any ERC-20 token as collateral, not just ether and its derivatives (The Block). It later raised $29 million in a Series A led by Pantera Capital, with Coinbase Ventures among the backers (The Block).
Karak and others are pushing multi-asset restaking as well, while EigenLayer has rebranded its broader ambitions as EigenCloud, casting restaking as the security base for a verifiable cloud of data availability, dispute resolution, and offchain compute. The contest is quietly shifting away from raw total value locked toward a harder question: which networks attract real demand from AVSs, and the fees that come with it?
What US regulators have, and have not, said
Staking spent years under a US legal cloud; the SEC’s 2023 enforcement action against Kraken’s staking-as-a-service product was the moment that put every operator on notice. The mood eased in 2025. On May 29, the SEC’s Division of Corporation Finance said certain protocol staking activities do not, on their own, amount to securities transactions (SEC). On August 5, staff extended similar comfort to liquid staking and the receipt tokens it produces (SEC).
Restakers should still read the fine print. Those statements stopped short of restaking itself, and LRTs that bundle AVS rewards, loyalty points, and active management sit in grayer territory. The guidance is also non-binding staff opinion rather than a rule, and SEC Commissioner Caroline Crenshaw dissented, cautioning that the reasoning could be stretched too far (SEC). The all-clear for staking, in short, does not automatically cover every product wearing the restaking label.
The bottom line for restakers
Restaking is one of the few genuinely new primitives crypto has produced in recent years, and the extra yield is real. So is the layered risk: smart-contract bugs, operator concentration, slashing on several fronts at once, bridge exploits like the Kelp DAO hack, and rules that are still being written. None of that makes restaking a bad idea, but it does make it a deliberate one.
For anyone weighing it up, a short checklist keeps the decision honest.
- Know exactly which AVSs your LRT is backing, and what their slashing conditions are.
- Check the audits, the operator set, and how concentrated that set really is.
- Watch the peg and on-chain liquidity, especially if you might need to exit in a hurry.
- Treat loyalty points as a bonus, not a reason to size up a position.
- Commit only what you could stand to lose in full, because correlated failures are the whole risk.
Used carefully, restaking turns idle security into productive capital. Used carelessly, it stacks risks faster than it stacks yield. The difference sits almost entirely in the homework.
By Marcus Vane, Senior Editor, HOGE Wire.