> ## Content Index
> Fetch the complete content index at: https://altcoininvestor.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# What Is Restaking? EigenLayer And AVS Explained
- URL: https://altcoininvestor.com/what-is-restaking/
- Published: 2026-09-12T21:04:47.000Z
- Updated: 2026-09-12T21:04:47.000Z
- Description: Restaking stacks AVS yield on ETH staking but adds slashing risk from every protocol you opt into. How the security-sharing mechanism works, real yields, and where the risk hides.
- Author: Gwen Harper
- Tags: Staking & Validation, DeFi Yield Strategies, Advanced, Liquid Staking

## What Restaking Is

![Validator keys connected to multiple security protocol layers through cryptographic chains](https://cdn.getmidnight.com/13448471d89a9cd8d7f71026a0334ec8/2026/09/ethereum-restaking-eigenlayer-after-h2-1.webp)

Restaking is a mechanism that allows you to reuse staked Ethereum to secure additional protocols called Actively Validated Services (AVS). Instead of staking ETH once for Ethereum consensus, restaking lets that same capital validate multiple services simultaneously. You earn additional yield from the AVS fees. You also accept additional slashing risk from every protocol you opt into.

Native Ethereum staking pays 3 to 4 percent APY and carries one set of slashing conditions: the Ethereum protocol rules. Restaking stacks AVS payments on top of that base yield and stacks the slashing conditions of each AVS on top of the Ethereum rules. A validator who opts into three AVS can be slashed for violating Ethereum consensus rules or for breaching the specific conditions of AVS 1, AVS 2, or AVS 3\. The slashing exposure compounds.

The core appeal is yield efficiency. Restaking turns idle staked capital into security for protocols that need validation but cannot afford to bootstrap their own validator sets. The core risk is slashing cascade. A bug, exploit, or governance failure in any one AVS can trigger slashing across thousands of restakers at once, because the same ETH secures many services.

EigenLayer launched the restaking category in 2023 and commands $15.258 billion in TVL as of mid-2026, representing 93.9 percent market share of Ethereum restaking. Symbiotic follows with roughly $1.6 billion, offering a permissionless modular architecture that accepts any ERC-20 as collateral. Karak manages about $102 million and supports multi-asset restaking including LP tokens, stablecoins, and wrapped Bitcoin. The protocols differ in collateral policies, slashing architecture, and governance model. All share the same fundamental structure: reuse staked capital to secure external services, earn fees, and accept layered slashing risk.

## How Restaking Works: The Mechanism

![Visualization of compound slashing risk when validator opts into multiple AVS protocols simultaneously](https://cdn.getmidnight.com/13448471d89a9cd8d7f71026a0334ec8/2026/09/ethereum-restaking-eigenlayer-after-h2-2.webp)

The restaking mechanism breaks into four components.

**1\. You stake ETH or deposit a liquid staking token.** EigenLayer accepts ETH and ETH-based liquid staking tokens like stETH, rETH, or cbETH. Symbiotic accepts any ERC-20 in isolated vaults. Karak accepts ETH, LSTs, LP tokens, stablecoins, and wBTC. You deposit collateral into the restaking protocol's smart contract.

**2\. You delegate to an Operator.** Operators are entities that run the infrastructure to validate AVS. Restakers delegate their staked capital to an Operator, who performs the actual validation work. The Operator receives a commission, typically 10 percent of AVS rewards in EigenLayer's case, and distributes the remainder to delegators. Symbiotic and Karak allow AVS to design their own operator payment structures.

**3\. The Operator opts into one or more AVS.** An AVS is any service that requires validation: a data availability layer, an oracle network, a rollup sequencer, a cross-chain bridge, a coprocessor. The AVS specifies its security requirements and slashing conditions. The Operator reviews those conditions and opts in on behalf of the delegators. Every time the Operator opts into a new AVS, the delegated capital secures that service and the delegators accept that service's slashing rules.

**4\. The AVS pays fees; slashing conditions activate.** The AVS pays the Operator in tokens, stablecoins, or protocol emissions. Those payments flow to delegators as restaking yield. Simultaneously, the AVS's slashing conditions become enforceable. If the Operator violates an AVS rule by going offline, signing conflicting messages, or failing a validity check, the AVS slashing mechanism can burn or redistribute a portion of the restaked collateral. The base Ethereum staking slashing conditions remain in force. The AVS slashing conditions stack on top.

You can think of restaking as a security marketplace. AVS rent validator capacity. Operators provide it. Restakers supply the capital and accept the risk. The mechanism works because bootstrapping a new validator set is expensive and slow, especially for smaller protocols. Restaking lets an AVS pay for shared security instead of building dedicated infrastructure. The trade-off is that the AVS does not control the validator set, and the validator set is exposed to correlated risk across many services.

EigenLayer introduced Unique Stake Allocation in April 2025 to isolate slashing risk. Under this model, a restaker can allocate specific portions of their stake to specific AVS, so a slashing event in one service does not automatically affect capital allocated elsewhere. Symbiotic builds isolation at the vault level, with each network defining its own slashing parameters. Karak uses a hybrid fraud-proof system called K2 to validate slashing claims before execution. These isolation mechanisms reduce but do not eliminate the correlation risk, because Operators still manage multiple AVS, and a compromised Operator can trigger slashing across all delegated capital.

If you want to understand how this compares to the base staking layer, [liquid staking tokens](https://altcoininvestor.com/what-are-liquid-staking-tokens/) offer a useful reference point. LSTs like stETH turn staked ETH into a tradable token that still earns base staking yield. Restaking takes that LST and reuses it for additional validation work. The layering is explicit: base staking, then liquid staking, then restaking. Each layer adds yield and adds risk.

## The Risk Stack: What Slashing Actually Means

![Calculating real restaking yield after factoring base staking rewards and AVS service fees](https://cdn.getmidnight.com/13448471d89a9cd8d7f71026a0334ec8/2026/09/ethereum-restaking-eigenlayer-after-h2-3.webp)

The primary risk in restaking is not market volatility. It is slashing. Slashing is the mechanism that enforces honest behavior in proof-of-stake systems. If a validator breaks a rule, a portion of their staked capital is destroyed or redistributed. Restaking multiplies the number of rules you can break.

Native Ethereum staking has three slashing conditions: double signing, surround voting, and extended downtime. These are well-defined and monitored. Slashing events on Ethereum are rare. The protocol has slashed fewer than 300 validators since the Beacon Chain launched.

Restaking adds the slashing conditions of every AVS you opt into. Those conditions vary widely. An oracle network might slash you for submitting incorrect price data. A data availability layer might slash you for failing to store blobs. A rollup might slash you for signing an invalid state transition. A bridge might slash you for attesting to a fraudulent cross-chain message. Each AVS defines its own rules, its own monitoring infrastructure, and its own slashing severity. You accept all of them when your Operator opts in.

The compound risk is what researchers call slashing cascade. A scenario where one AVS bug, exploit, or governance attack triggers slashing across many restakers simultaneously. This is not theoretical. In April 2026, Kelp DAO suffered a $300 million exploit when attackers linked to the Lazarus Group compromised its LayerZero bridge. The exploit did not trigger slashing, but it triggered $5.4 billion in withdrawals across the restaking sector. The mechanism held. User confidence did not. A slashing event would have had a larger impact, because slashed capital cannot be withdrawn. It is burned or redistributed.

Correlation is the core problem. Restaking works by sharing the same capital across multiple services. If those services share dependencies like a common oracle, a common Operator set, or a common governance token, a failure in one can propagate to others. Increasing the number of AVS reduces individual risk exposure to any one service, but only marginally, because the slashing conditions are not independent. A compromised Operator, a shared infrastructure bug, or a governance takeover can affect many AVS at once.

EigenLayer's redistributable slashing model, launched in July 2025, allows slashed assets to be redirected to a penalty vault or redistributed to non-slashed delegators rather than burned outright. This reduces the total loss but does not eliminate it. Symbiotic's isolated vault architecture keeps different collateral pools separate, so a slashing event in one AVS cannot directly affect capital in a vault securing a different AVS. Karak's fraud-proof system requires evidence of an infraction before slashing executes, adding a verification layer. These designs mitigate risk. They do not remove the fundamental exposure: you are securing multiple services with overlapping failure modes.

The second-order risk is Operator concentration. As of mid-2026, a small number of large Operators manage the majority of restaked ETH. If one major Operator is compromised, goes offline, or makes a configuration error, thousands of delegators are slashed simultaneously. This is a governance and decentralization problem, not a technical one. The restaking protocols have no mechanism to limit Operator market share, and delegators optimize for yield, not decentralization. The result is a validator set more concentrated than base Ethereum staking.

A third risk is toxic yield. AVS that offer higher payments often do so because they impose more aggressive slashing conditions, have less mature infrastructure, or carry higher technical risk. Restakers chasing yield are incentivized to opt into riskier services. This creates a race to the bottom where the protocols offering the highest APY are also the ones most likely to trigger a slashing event. The mechanism does not penalize this behavior until the slashing event occurs, at which point the capital is already gone.

No major slashing event has occurred across any restaking protocol as of mid-2026\. The slashing mechanisms remain largely untested at scale. This is not reassuring. It means the risk models are based on simulations and assumptions, not observed outcomes. When the first large-scale slashing event happens, the market will reprice restaking risk. Until then, the yield looks attractive and the slashing looks hypothetical.

For context on risk layering in DeFi, [impermanent loss](https://altcoininvestor.com/what-is-impermanent-loss/) offers a useful parallel. IL is a yield mechanism where the risk is hidden in the position structure, not the market price. Restaking works the same way. The risk is in the slashing conditions you accepted when you delegated, not in the daily APY you observe.

## Where Restaking Yield Actually Comes From

Restaking yield decomposes into three components: base staking rewards, AVS service fees, and token incentives. The composition matters because each component has a different risk profile and a different sustainability model.

**Base staking rewards.** If you restake native ETH or an LST, you continue earning Ethereum staking yield, currently 3 to 4 percent APY. This component is stable, protocol-guaranteed, and independent of the restaking layer. It accrues as long as your validator remains active and does not get slashed by Ethereum consensus rules. This is the same yield you would earn from [staking ETH directly](https://altcoininvestor.com/best-crypto-staking-platform/). Restaking does not increase it. It stacks additional yield on top of it.

**AVS service fees.** AVS pay Operators to secure their networks. Those fees are denominated in the AVS's native token, in stablecoins, or in ETH. The Operator takes a commission, typically 10 percent in EigenLayer's model, and distributes the remainder to delegators. AVS fees are the real restaking yield. They come from protocol revenue, user transaction fees, or treasury reserves. The sustainability depends on whether the AVS generates enough revenue to pay for security. If the AVS has no revenue and funds fees from token emissions, the yield is dilutive and unsustainable. If the AVS charges users and shares that revenue with validators, the yield is real.

As of mid-2026, most AVS are paying fees from token emissions or points programs, not from user revenue. This is early-stage behavior. The protocols are subsidizing yield to attract validators. The yield is real in the sense that you receive tokens, but it is not sustainable unless the AVS transitions to a fee-paying user base. The historical precedent is liquidity mining in DeFi. High APY during the bootstrap phase, sharp drop once emissions taper and the protocol must rely on organic revenue. Expect the same pattern here.

AVS fees in the 1 to 3 percent range are realistic when the service has product-market fit and charges users. Fees above 5 percent typically indicate heavy emissions or unsustainable subsidy programs. The current restaking yield stack lands in the 4 to 7 percent range when you add base staking plus AVS fees. The lower end of that range is durable. The upper end is not.

**Token incentives and points programs.** Many restaking protocols and LRTs issue points, governance tokens, or protocol-specific rewards on top of AVS fees. Kelp DAO's rsETH, for example, offered Kelp Miles that stacked on top of EigenLayer points during the 2024-2025 restaking boom. Ether.fi, which commands 65 percent of the liquid restaking token market with $7.83 billion in TVL, distributes its ETHFI governance token to restakers. These programs are marketing, not yield mechanisms. They are designed to attract TVL during the growth phase. They end when the protocol matures or when the treasury runs out of tokens to distribute.

Points programs are especially opaque. The conversion rate from points to tokens is unknown until the airdrop occurs. The value of the token is unknown until it trades. The APY displayed on dashboards during the points phase is speculative. You are farming an unknown quantity of an unknown asset. This is not yield. It is lottery-ticket speculation dressed up as income. Some users made significant returns from EigenLayer and LRT airdrops in 2024-2025\. That does not make the mechanism repeatable or the yield predictable.

When evaluating restaking yield, subtract the points component entirely. Focus on base staking plus AVS fees. If the combined real yield does not justify the slashing risk, the position is not worth taking. If you are restaking because of an expected airdrop, you are speculating, not earning. The distinction matters for risk management and tax treatment.

For yield decomposition principles, [LRT vs LST](https://altcoininvestor.com/lrt-vs-lst/) walks through how liquid restaking tokens stack on top of liquid staking tokens and where the incremental yield comes from. The same decomposition logic applies here: each layer of yield comes with a corresponding layer of risk. If you cannot name the risk, do not take the position.

## EigenLayer, Symbiotic, and Karak: What Each Protocol Does Differently

EigenLayer rebranded to EigenCloud in June 2025 and markets itself as a verifiable cloud platform. EigenLayer is the restaking security layer. EigenDA, EigenCompute, and EigenAI are the products built on top. The protocol accepts only ETH and ETH-based liquid staking tokens as collateral. It operates a curated AVS marketplace where services are vetted before being offered to restakers. EigenLayer uses Unique Stake Allocation to let restakers isolate portions of their capital to specific AVS, reducing but not eliminating slashing correlation. Slashing went live in April 2025 with operator sets and unique stake. Redistributable slashing followed in July 2025, allowing slashed assets to be redirected rather than burned.

EigenLayer's design philosophy is institutional and Ethereum-native. The curated AVS list provides some quality control. The ETH-only collateral model keeps the protocol aligned with Ethereum security assumptions. The trade-off is permissioned access and limited collateral flexibility. If you want to restake non-ETH assets or launch an AVS without going through EigenLayer's vetting process, you cannot use EigenLayer.

Symbiotic launched as a permissionless alternative. It accepts any ERC-20 as collateral in isolated per-network vaults. Each vault has its own collateral type, its own slashing rules, and its own AVS. The core contracts are immutable. There is no governance token as of mid-2026, and the protocol has no central curation mechanism. Anyone can launch an AVS on Symbiotic. Anyone can create a vault. The architecture is modular. Restakers choose which vault to deposit into, and each vault exposes them only to the slashing conditions of the AVS that vault secures. If an AVS in vault A fails, your capital in vault B is unaffected.

Symbiotic's design philosophy is flexibility and composability. The downside is that permissionless systems attract low-quality projects. There is no filtering mechanism. Restakers must evaluate AVS risk themselves. The collateral flexibility means you can restake assets with poor liquidity or unstable pegs, increasing the risk of cascading liquidations in a market downturn. Symbiotic trades safety for configurability.

Karak operates a multi-asset restaking protocol with $102 million in TVL across 29,000 ETH, representing 0.6 percent market share. Beyond ETH and LSTs, Karak accepts LP tokens, stablecoins, and wrapped Bitcoin as restaking collateral. The protocol uses a hybrid security model called K2, which combines Ethereum consensus with its own fraud-proof system. K2 validates slashing claims before executing them, adding a verification layer that reduces false positives. Karak supports generalized restaking across Layer 2 ecosystems including Arbitrum and Base, not just Ethereum mainnet.

Karak's design philosophy is multi-chain expansion and broader collateral support. The trade-off is complexity and a less-proven security model. K2 is new. The fraud-proof system has not been tested under adversarial conditions. Accepting LP tokens as collateral introduces impermanent loss risk on top of slashing risk. Accepting stablecoins introduces depeg risk. The capital efficiency is higher, but so is the tail risk.

If your priority is liquidity, ecosystem maturity, and alignment with Ethereum's security assumptions, use EigenLayer. If you need collateral flexibility, vault-level isolation, and permissionless AVS deployment, evaluate Symbiotic. If you want to restake non-ETH assets or secure Layer 2 protocols, Karak is the only option. All three share the same fundamental risk: you are securing multiple services with the same capital, and a failure in any one service can cost you principal.

## Real Yield vs Speculative Yield: What You Are Actually Earning

The difference between real yield and speculative yield in restaking is the difference between protocol-generated fees and token incentives. Real yield comes from users paying for a service. Speculative yield comes from a treasury distributing tokens to attract TVL. One is sustainable. The other is not.

As of mid-2026, most restaking yield is speculative. AVS are subsidizing payments to bootstrap validator sets. Points programs dominate the LRT landscape. Governance token emissions are high. User-generated fee revenue is low. This is normal for an early-stage market. It is also unsustainable. When the subsidies end, the yield will drop. Restakers who entered for the high APY will leave. TVL will contract. The protocols that survive will be the ones with real revenue.

To evaluate whether an AVS is paying real yield, check where the fees come from. Does the AVS charge users for data availability, oracle updates, sequencing, or other services? Does that revenue exceed the cost of paying validators? If the answer is yes, the yield is real. If the AVS has no revenue and pays validators from token emissions or a reserve fund, the yield is temporary.

EigenDA, EigenLayer's data availability service, is one of the few AVS with observable user revenue. Rollups pay EigenDA to store blobs. That revenue flows to restakers as AVS fees. The mechanism is sustainable as long as rollups continue using the service and paying for it. Other AVS in the EigenLayer ecosystem are still in the subsidy phase. Symbiotic and Karak have even less fee-generating activity, because most of their AVS are experimental or have not launched yet.

The realistic durable yield from restaking in mid-2026 is 4 to 7 percent APY. That is base Ethereum staking at 3 to 4 percent plus AVS fees in the 1 to 3 percent range. Higher yields exist, but they are subsidized and will compress. Lower yields are possible if you opt into AVS with minimal fee generation. If you see a restaking opportunity advertising 15 percent APY, decompose it. The excess above 7 percent is coming from points, emissions, or a reserve fund that will run out.

Kelp DAO's rsETH is a case study in speculative yield. Kelp issued Kelp Miles on top of EigenLayer points and pulled significant TVL during the 2024-2025 boom. The implied APY from points was high. The actual yield once the points converted to tokens was lower. The Kelp exploit in April 2026 triggered $5.4 billion in withdrawals and collapsed the rsETH peg. The protocol survived, but the lesson is clear: points are not yield, and high TVL during a points program does not indicate protocol health.

For a broader view of how yield stacking works in DeFi, the [crypto yield calculator](https://altcoininvestor.com/crypto-yield-calculator/) lets you model entry costs, hold periods, and net returns after fees. Restaking yield should be modeled the same way. Subtract operator commissions. Subtract gas costs for deposits and withdrawals. Subtract the expected value of slashing risk. What remains is your net return. If that number is below what you can earn from [base staking](https://altcoininvestor.com/staking-rewards-calculator/), restaking is not worth the additional complexity and risk.

## When Restaking Makes Sense, When It Doesn't

Restaking makes sense when you understand the slashing conditions of every AVS your Operator has opted into, when the incremental yield justifies the incremental risk, and when you have a plan for monitoring operator behavior and AVS health. It does not make sense when you are chasing APY without decomposing the source, when you cannot name the failure modes, or when you are restaking because everyone else is.

The decision framework is simple. Start with your base case: staking ETH directly or holding an LST. That position earns 3 to 4 percent with one set of slashing conditions. Restaking adds 1 to 3 percent in AVS fees and adds the slashing conditions of every AVS you secure. Is the extra 1 to 3 percent worth the extra slashing exposure? If yes, restaking makes sense. If no, it does not.

The answer depends on your risk tolerance, your capital size, and your ability to monitor the position. Large institutional holders with dedicated risk teams can evaluate AVS slashing conditions, monitor operator performance, and exit positions quickly if conditions deteriorate. Retail holders without those resources are accepting risk they cannot measure. The information asymmetry is large. Operators know which AVS they have opted into and what the slashing conditions are. Delegators often do not, because the information is spread across protocol documentation, governance forums, and contract state.

If you are going to restake, use a protocol with transparent slashing conditions and vault-level isolation. EigenLayer publishes AVS documentation. Symbiotic isolates risk at the vault level. Karak's fraud-proof system adds a verification step before slashing. All three are better than restaking on a protocol with opaque slashing rules or a single pooled collateral model where one AVS failure affects everyone.

Avoid restaking if you are doing it for points. Points programs are marketing. The conversion rate is unknown. The token value is unknown. You are speculating, not earning. If the protocol you are evaluating advertises restaking yield in double digits and the breakdown shows most of it coming from points, skip it. The real yield is in the low single digits. The rest is hype.

Avoid restaking if the AVS your Operator has opted into are unaudited, have unclear slashing conditions, or are controlled by a small governance token holder set. These are the AVS most likely to trigger a slashing event, either through a bug, an exploit, or a governance attack. The yield is not worth the risk of losing principal.

Use restaking if you have a long time horizon, if you are comfortable with the slashing risk, and if the AVS you are securing have real user revenue and sustainable fee structures. Use it if you can monitor operator performance and AVS health on-chain. Use it if the incremental 1 to 3 percent APY materially improves your portfolio return and you have sized the position so that a total loss would not affect your broader strategy.

Do not use restaking if you are new to DeFi, if you do not understand slashing, or if you cannot read smart contracts. The risk is in the details. The details are in the contracts. If you cannot verify the claims, you cannot manage the risk.

## The Takeaway

Restaking is a security-sharing mechanism that lets you reuse staked ETH to validate additional services called AVS. You earn base staking yield plus AVS fees, typically 4 to 7 percent APY. You accept base Ethereum slashing conditions plus the slashing conditions of every AVS your Operator opts into. The yield stacks. The risk stacks. The mechanism is efficient when AVS have real user revenue and pay validators from fees. It is speculative when AVS pay from token emissions or points programs.

EigenLayer dominates with $15.258 billion in TVL and a curated, Ethereum-native approach. Symbiotic offers permissionless deployment and multi-collateral flexibility. Karak supports multi-asset restaking across Layer 2 ecosystems. All three expose you to slashing cascade risk, operator concentration, and toxic yield if you opt into high-risk AVS.

The core question is whether the incremental 1 to 3 percent yield justifies the incremental slashing exposure. For large holders with risk infrastructure, it might. For retail holders chasing points, it does not. The first major slashing event will reprice the entire market. Until then, the yield looks attractive and the risk looks hypothetical. It is not hypothetical. It is deferred. The slashing conditions are live. The mechanisms are untested at scale. Decompose the yield, verify the slashing rules, and size the position accordingly.

## Frequently Asked Questions

### What is the difference between staking and restaking?

Staking locks ETH to secure Ethereum consensus and earns 3 to 4 percent APY with one set of slashing conditions. Restaking reuses that staked ETH to secure additional services called AVS, earning extra yield but accepting additional slashing risk from each AVS. The same capital secures multiple protocols, stacking both the income and the failure modes.

### How much can you earn from restaking?

Realistic restaking yield in mid-2026 is 4 to 7 percent APY. That decomposes into 3 to 4 percent from base Ethereum staking plus 1 to 3 percent from AVS service fees. Higher advertised yields typically come from points programs or token emissions, which are speculative and unsustainable. Real yield comes from AVS that charge users and share revenue with validators.

### What is slashing in restaking?

Slashing is the penalty mechanism that destroys or redistributes staked capital when a validator breaks protocol rules. In restaking, you accept the slashing conditions of Ethereum plus the slashing conditions of every AVS your operator opts into. A violation in any one service can trigger slashing. The risk compounds because the same capital secures multiple protocols with overlapping failure modes.

### What is an AVS in EigenLayer?

An AVS, or Actively Validated Service, is any protocol that requires validation but does not have its own validator set. Examples include data availability layers, oracle networks, rollup sequencers, and cross-chain bridges. AVS pay restakers to secure their services. In return, restakers accept the AVS slashing conditions and earn fees on top of base staking yield.

### Is restaking worth the risk?

Restaking is worth the risk if you understand the slashing conditions of every AVS you secure, if the incremental 1 to 3 percent yield justifies the added exposure, and if you can monitor operator and AVS health on-chain. It is not worth it if you are chasing points, cannot verify slashing rules, or are new to DeFi. The first major slashing event will reprice the market. Until then, the risk is real but untested at scale.

The Weekly Yield Report

You just saw how three restaking protocols decompose layered yield and slashing risk. Both the APY and the failure modes will shift as AVS deployments mature.

Every Thursday: where crypto yield actually is - stablecoins, liquid staking and DeFi lending, with the risk named next to the rate and what changed since last week.

[Get it free every Thursday](#/portal/signup) 

Free. No trade calls, no allocations, no hype. Unsubscribe in one click.