Ethereum Restaking Explained: EigenLayer, LRTs, and the $18B Ecosystem (2026)
How EigenLayer created a new yield layer on top of Ethereum staking, which liquid restaking tokens dominate the $18B market, what the Kelp DAO $293M exploit revealed about LRT risk, and how to evaluate restaking protocols.
EigenLayer has grown into an $18 billion restaking ecosystem that allows staked ETH to secure additional services beyond the Ethereum consensus layer. It commands 93.9% of the restaking market, with Symbiotic ($897M) and Karak ($102M) as distant competitors. The protocol describes itself as a "verifiable cloud" — decentralized compute infrastructure secured by restaked ETH.
Liquid restaking tokens (LRTs) — led by Ether.fi ($5.6B TVL) and Kelp DAO ($2B+) — abstract the complexity of restaking into single-token deposits. But the April 2026 Kelp DAO exploit ($293M, Lazarus Group) demonstrated that LRT bridge infrastructure remains a critical vulnerability: the exploit triggered a $13 billion DeFi-wide drawdown.
This guide covers how restaking works, every major LRT protocol, the AVS ecosystem, the competitive landscape, and the specific risks that make restaking fundamentally different from base liquid staking.
What Is Restaking?
Standard Ethereum staking secures one thing: the Ethereum consensus layer. Validators attest to blocks and earn ~3-4% APY for doing so. Restaking extends this model: ETH that is already staked on the beacon chain can simultaneously be used as security collateral for additional decentralized services.
The key insight is that most staked ETH sits idle between attestation duties. Restaking puts that idle capital to work by allowing validators to opt in to securing Actively Validated Services (AVS) — oracle networks, data availability layers, bridges, sequencers, and keeper networks — earning supplemental fees from each AVS they secure.
The trade-off is additional slashing surface. A restaked validator can be penalized by both Ethereum (for consensus faults) and by any AVS it has opted into (for service-specific misbehavior). This layered slashing risk is what makes restaking fundamentally different from — and riskier than — base staking.
EigenLayer: The Restaking Infrastructure
Launched in 2023, EigenLayer created the restaking category. As of mid-2026, it held approximately $15-18 billion in restaked ETH across 4.36 million ETH, reaching an all-time high TVL of $19.7 billion. The protocol commands 93.9% of the total restaking market.
How it works: Stakers deposit ETH (or LSTs like stETH) into EigenLayer smart contracts. They then delegate to operators — professional node runners who register with specific AVS. Each AVS defines its own slashing conditions and reward mechanisms. The operator runs the AVS software alongside their Ethereum validator client and earns fees from the AVS, which flow through to delegators.
EigenDA — the flagship AVS: EigenDA is a data availability layer that allows rollups to post data to a decentralized network of EigenLayer operators instead of Ethereum calldata. It remains the single largest consumer of restaked security and the clearest demonstration of the restaking thesis.
The "verifiable cloud" vision: EigenLayer has evolved beyond simple restaking into what the team calls a "verifiable cloud" — targeting enterprise applications and institutional users who need decentralized compute infrastructure. Vertical AVS specialization is reshaping how restaking generates value, with specialized services for AI inference verification, cross-chain state proofs, and MEV supply chain components.
The Competitive Landscape
| Protocol | TVL | ETH Restaked | Market Share |
|---|---|---|---|
| EigenLayer | $15.3-18B | 4,364,467 | 93.9% |
| Symbiotic | $897M | 256,533 | 5.5% |
| Karak | $102M | 29,055 | 0.6% |
Symbiotic: The primary challenger, launched by former Lido contributors with a modular design that allows any ERC-20 token (not just ETH/LSTs) to be used as restaking collateral. At $897M TVL and 5.5% market share, Symbiotic has demonstrated that alternative restaking designs can attract meaningful capital, though it remains a fraction of EigenLayer's scale.
Karak: Focused on multi-chain restaking with support for assets beyond Ethereum. At $102M TVL (0.6% share), Karak is early-stage but differentiated by its cross-chain approach.
Liquid Restaking Tokens (LRTs): The Access Layer
Direct restaking through EigenLayer requires selecting operators, choosing AVS, and managing delegation — a complexity barrier for most users. Liquid restaking tokens solve this by abstracting the entire process into a single deposit action: send ETH, receive a tradeable LRT that earns both staking and restaking yield automatically.
| Protocol | Token | TVL | Key Feature | Notable Event |
|---|---|---|---|---|
| Ether.fi | eETH / weETH | $5.6B | Non-custodial keys, auto-AVS | Market leader since 2024 |
| Kelp DAO | rsETH | $2B+ | Multi-LST restaking, 10+ L2s | $293M exploit Apr 2026 |
| Swell | rswETH | $292M | Discontinued own L2 Jun 2026 | 859M SWELL burn Apr 2026 |
| Bedrock | uniETH | $24M | Multi-asset (uniETH + uniBTC) | $2M uniBTC exploit 2024 |
Ether.fi (eETH/weETH) — The Market Leader
Ether.fi leads the LRT category with $5.6 billion in TVL. Users deposit ETH and receive eETH (rebasing) or weETH (non-rebasing, optimized for DeFi composability on Aave, Pendle, and lending protocols). Behind the scenes, Ether.fi handles staking, EigenLayer delegation, and AVS selection automatically. Unlike most LRT protocols, Ether.fi allows stakers to retain control of their validator keys — a meaningful non-custodial distinction.
Kelp DAO (rsETH) — Recovery After the $293M Exploit
Kelp DAO was the second-largest LRT protocol before the April 18, 2026 exploit that shook the entire DeFi ecosystem. rsETH accepts multiple LSTs (stETH, rETH, ETHx) for restaking through EigenLayer, and is deployed across 10+ Layer 2 networks and 40+ DeFi platforms.
Restaking Risks: What Makes This Different
Restaking introduces risk layers that do not exist in base liquid staking. Understanding these is essential before depositing into any LRT.
Layered smart contract risk: An LRT deposit touches at minimum three smart contract systems: the LRT protocol itself, EigenLayer's core contracts, and the AVS contracts being secured. A bug in any layer can result in fund loss. Base LSTs like stETH only touch one protocol's contracts.
Expanded slashing surface: Restaked validators can be slashed by both Ethereum (for consensus faults) and by each AVS they secure (for service-specific misbehavior). Correlated slashing — where multiple validators are penalized simultaneously — amplifies losses. The more AVS a validator opts into, the more slashing vectors exist.
Bridge risk: Most LRTs are available across multiple L2 networks via bridges. As the Kelp DAO exploit proved, bridge infrastructure is the weakest link — not the core restaking contracts. A 1-of-1 bridge multisig is a single point of failure regardless of how decentralized the underlying protocol is.
Systemic contagion: With $18 billion in restaked ETH, EigenLayer is now "too big to fail" infrastructure. A critical vulnerability in EigenLayer's core contracts could cascade across every LRT, every AVS, and every DeFi protocol that accepts LRTs as collateral. This systemic risk does not exist for base LSTs, which only depend on the Ethereum consensus layer.
Operator concentration: While restaking appears decentralized because many users participate, actual operator control clusters around a few professional infrastructure providers. This creates correlated risk — if a major operator misconfigures their setup, all delegators to that operator are affected simultaneously.
The Bottom Line
Restaking represents the most significant innovation in Ethereum's staking economy since liquid staking itself. EigenLayer's $18 billion TVL and the emergence of a verifiable cloud vision demonstrate genuine product-market fit. But the category is younger, less battle-tested, and structurally more complex than base liquid staking.
The April 2026 Kelp DAO exploit crystallized what was previously theoretical: bridge infrastructure in the LRT stack is a critical vulnerability that can cause hundreds of millions in losses and trigger DeFi-wide contagion. The distinction between a protocol's core contracts (which were not breached) and its peripheral infrastructure (which was) is cold comfort when $293 million is minted from thin air.
For users evaluating restaking, the decision framework is clear: Ether.fi (eETH/weETH) is the market leader with the deepest DeFi liquidity and longest LRT track record. EigenLayer itself has not been exploited. But every LRT adds smart contract layers, slashing vectors, and bridge dependencies on top of base staking risk. The supplemental yield from restaking is payment for accepting those additional risks — whether that payment is adequate is a judgment each staker makes with their own capital.