Restaking Explained: Yield, Slashing and Systemic Risk
Restaking lets staked ETH and Bitcoin secure extra networks for extra yield, led by EigenLayer. With slashing now live and the SEC weighing in, the risks are no longer theoretical.
Staking once told a simple story: lock up Ether, help run a validator, collect a yield, and that was the end of it. Restaking breaks that one-to-one relationship. It lets the same staked capital back more than one system at the same time, turning Ethereum’s deep pool of staked ETH into shared economic security that other protocols can rent. The concept jumped from a 2023 design paper to one of the biggest experiments in decentralized finance, and in 2026 it is finally being stress-tested with real penalties and real attention from regulators in Washington.
What restaking actually means
Plain staking commits a token to secure a single network. Restaking takes capital that is already staked, for example Ether sitting behind a validator or a liquid staking token such as Lido’s stETH, and pledges it a second time as collateral for extra services. In exchange for accepting more rules and more ways to lose money, the restaker earns more rewards. EigenLayer, the protocol that coined the term, frames restaking as a marketplace where holders of staked assets supply security and new applications buy it, instead of every application bootstrapping a fresh validator set from scratch, as its own documentation lays out.
The pitch is capital efficiency. A new oracle, bridge, or data availability layer normally has to attract its own stakers and convince the market that its token is worth enough to deter attacks. Restaking lets that service borrow trust already parked behind Ethereum, which lowers the cost of launching infrastructure that lives or dies on credibility.
How restaking works: operators, services and penalties
Three roles keep the machine turning. Restakers supply the capital. Operators run the actual node software and take delegated stake, much like a staking pool does today. The services that buy security are called Actively Validated Services, or AVSs, and they include oracles, cross-chain bridges, rollup sequencers, and data availability layers. Each AVS writes its own slashing conditions, the specific offenses that let it confiscate part of an operator’s stake. A restaker who delegates to an operator inherits exposure to every AVS that operator has joined, so picking an operator is itself a risk decision.
- Restakers supply staked assets and choose which operators to back.
- Operators run the software for one or more services and put delegated stake at risk.
- Actively Validated Services (AVSs) buy security and define the rules that trigger slashing.
In practice most restakers never run anything themselves. They deposit into a liquid restaking protocol or delegate to a professional operator, which keeps the experience close to ordinary staking even though the risk profile underneath is markedly different. That gap, between how simple it feels and how much is actually moving, is part of what makes the sector worth watching closely.
EigenLayer and the protocol that defined the category
Restaking is largely the story of EigenLayer, built by a team led by Sreeram Kannan that turned the idea into working code. A phased mainnet rollout through 2023 and 2024, paired with a points program that hinted at a future airdrop, pulled in deposits at a startling pace. Total value locked climbed to an all-time high near $19.7 billion before the points frenzy cooled, and EigenLayer still holds the large majority of all restaked value, with market trackers putting the wider restaking sector around $16 billion. The project has since rebranded its broader stack as EigenCloud.
The EIGEN token, launched in 2024 as a governance and security asset, has had a rough ride. It traded near $0.21 in mid-2026, a small fraction of its debut value, after sliding roughly 91% across 2025 as token emissions outran real demand, CoinDesk reported; CoinGecko pegs its market capitalization near $158 million. In December 2025 the Eigen Foundation proposed steering rewards toward participants who actively secure services rather than passive holders, and routing fees from AVSs and cloud products back to holders through buybacks.
The flagship service built on the protocol is EigenDA, a data availability layer that rollups can use to post transaction data more cheaply than writing it straight to Ethereum. It is the clearest case of the model working as intended: a real product paying for security it did not have to bootstrap alone. The open question is how many other AVSs can show the same durable demand once subsidies fade.
Liquid restaking tokens and the yield layer
Locked collateral is dead weight, so a second industry sprang up to free it. Liquid restaking tokens, or LRTs, work like liquid staking tokens one floor higher: a user deposits ETH or an LST, the protocol restakes it through EigenLayer or a rival, and mints a tradable receipt that keeps earning while it circulates through DeFi. ether.fi’s weETH, Renzo’s ezETH, Puffer’s pufETH, and Kelp DAO’s rsETH are among the largest. DeFiLlama tracks liquid restaking protocols at roughly $7 billion in combined value, well below the manic peaks of 2024.
| Platform | Native asset | Token | Approach |
|---|---|---|---|
| EigenLayer | ETH and LSTs | EIGEN | Original restaking marketplace for AVSs |
| ether.fi | ETH | weETH | Largest liquid restaking issuer |
| Renzo | ETH and LSTs | ezETH | Liquid restaking with cross-chain reach |
| Puffer | ETH | pufETH | Liquid restaking with anti-slashing tooling |
| Kelp DAO | ETH and LSTs | rsETH | Multi-asset liquid restaking |
| Symbiotic | Any ERC-20 | None yet | Collateral-agnostic shared security |
| Babylon | Bitcoin | BABY | Native Bitcoin staking, no bridge |
Figures move daily; treat platform totals as snapshots rather than fixed values.
These receipts can be looped, lent, and posted as collateral elsewhere, which multiplies yield and risk in equal measure. A token two or three layers removed from the underlying ETH can wobble if any layer breaks, as several brief LRT depegs in 2024 demonstrated.
Beyond Ethereum: Symbiotic, Karak and Bitcoin via Babylon
EigenLayer no longer has the field to itself. Symbiotic, live on Ethereum mainnet since 28 January 2025, takes a more permissionless, collateral-agnostic line: almost any ERC-20 can serve as security, and networks assemble their own risk parameters rather than accept a single template. It has gathered more than a billion dollars in deposits and partnerships with names such as Chainlink and Ethena. Karak provides yet another restaking venue with its own approach to supported assets.
The larger frontier may be Bitcoin. Babylon lets holders stake BTC to secure other chains without wrapping or bridging the coins, using native Bitcoin scripts and time-locks so funds never leave the base layer. The protocol pulled in well over a billion dollars of Bitcoin within months of launch, Blockworks noted, and spent 2025 shipping multi-staking, where one BTC position can secure several networks at once, alongside an EVM-compatible layer. Babylon calls 2025 the year idle Bitcoin finally found a yield-bearing job beyond cold storage.
Slashing goes live, the test restaking was waiting for
For most of its life, restaking offered the reward half of proof-of-stake without the punishment half. That changed on 17 April 2025, when EigenLayer switched on slashing, the feature its founders had always called central to the design, according to CoinDesk. Operators that misbehave can now lose stake for real, not just in theory.
Two ideas make the new system workable. Unique Stake lets an operator earmark a slice of security so that only one service can slash it, which stops the same dollar from being booked as protection for many services at once. Operator Sets group operators under a given service’s rules. EigenLayer later added redistributable slashing through ELIP-006, so confiscated funds can be routed to a harmed party instead of simply burned, though native ETH and EIGEN are not yet eligible. The team spelled out the mechanics on its engineering blog.
The risks: rehypothecation, looping and Vitalik’s warning
Restaking concentrates risk in ways that are easy to wave away during a bull run. The same collateral can sit behind many services at once, a form of rehypothecation that lifts yield until one bad event triggers correlated losses across everything that leaned on it. Liquid restaking stacks leverage on top, and elaborate slashing rules raise the odds of bugs and of fights over whether a penalty was actually earned.
Ethereum co-founder Vitalik Buterin flagged the deeper hazard back in May 2023, before the boom. In an essay titled Don’t overload Ethereum’s consensus, he argued that reusing validator stake for extra jobs is mostly fine, but trying to recruit Ethereum’s social consensus to rescue an application is not. If a restaked service blows up badly enough, the fear is that its backers lobby for a chain-level bailout, dragging outside disputes into the base protocol. Each such hook, he warned, makes the core itself more fragile.
There is plainer operational risk too. Restaking routes billions through a young stack of smart contracts, and a large share of operating power tends to cluster among a handful of professional operators, which cuts against the decentralization the whole design is meant to reinforce.
Where US regulators stand
The shift is striking given the SEC’s earlier posture. In 2023 the agency forced Kraken to shut its US staking-as-a-service program and pay a $30 million penalty, and it named staking in enforcement cases against major exchanges. Then the mood changed. On 29 May 2025 the SEC’s Division of Corporation Finance stated that common forms of protocol staking, spanning solo, self-custodial and custodial arrangements, do not by themselves count as securities transactions. On 5 August the staff extended similar reasoning to certain liquid staking activities, saying staking receipt tokens are not offered or sold as securities.
The relief comes wrapped in caveats. These are staff views, not formal rules; they bind no one, and they do not foreclose later enforcement or a different stance from the full Commission. Commissioner Caroline Crenshaw dissented, arguing the statements read the law far too generously. Restaking clouds the picture again, because AVS rewards are paid for an additional service, the very kind of managerial effort that securities tests probe. None of the 2025 guidance squarely tackles that wrinkle.
What to watch in the second half of 2026
The next phase is about whether restaking earns its keep. A few signals are worth tracking:
- Whether AVSs pay real fees rather than points and emissions, now that EigenLayer’s proposed Incentives Committee aims to reward productive stake.
- The first high-profile slashing events, which will show how cleanly penalties and redistribution work under pressure.
- Babylon’s multi-staking and EVM push, and whether Bitcoin restaking can close the gap on Ethereum.
- Consolidation among liquid restaking tokens as yields compress and the weakest receipts lose liquidity.
- Any move by the full SEC, or by Congress, to turn 2025’s staff guidance into something binding.
Restaking has already proven it can pull in capital. The harder question, now that the penalties are switched on and regulators are watching, is whether shared security can survive its first serious failure without dragging down the chains it was built to protect.
By the HOGE Wire desk, reporting on DeFi and on-chain markets.