FinWiki
日EN中

Restaking and AVS Landscape Matrix · EigenLayer vs Symbiotic vs Karak vs Mellow vs EtherFi vs Lido CSM

Confidence: Likely Updated 2026-05-26 Review by 2026-11-25 Sources 13 Machine-translated Original (JA)
#systems#matrix#restaking#eigenlayer#symbiotic#karak
On this page

TL;DR

  • Compare 6 restaking infrastructures across 2026-Q2 TVL · restaking model · slashing · AVS economics · correlation risk · operator concentration · governance · audit · Lido overlap · regulatory exposure on 10 axes
  • The 3 quadrants of restaking model: Operator-delegated (EigenLayer · Karak · Symbiotic) · LRT vault aggregator (EtherFi · Mellow · Renzo) · Native protocol restaking (Lido CSM · LST itself provides the restake hook)
  • TVL 2026-Q2 estimate: EigenLayer ~$14B · Symbiotic ~$3.2B · EtherFi LRT (eETH) ~$5.5B · Mellow LRT ~$1.8B · Karak ~$1.0B · Lido CSM ~$280M (early)
  • AVS landscape, 5 large names: EigenDA (~$120M revenue 2026 estimate · multiple ZK rollup users) · Hyperlane (borrows EigenLayer/Symbiotic through ISM) · Espresso (shared sequencer · Polygon/zkSync/Caldera and others connected) · Lagrange (coprocessor + state proof) · Witness Chain (POI · physical-infrastructure verification)
  • 2025 年 EigenLayer slashing formally enabled · actual slash events ≤ 5 件 · cumulative < $5M (far below early concerns) · but correlation risk (rehypothecation) is still not sufficiently priced
  • Lido stake overlap rate: ~25% of EigenLayer TVL comes from stETH restaking · ~40% in Symbiotic · Mellow is entirely LST-set based. Lido is the “base-layer liquidity source” of the restaking ecosystem
  • Routing: eigenlayer-overview (core) · liquid-staking-restaking-cex-exposure (LRT CEX exposure) · Pectra EIP-7251 · 2048 ETH validator (institutional staking route)

Wiki route

This entry sits under systems index. Read it against EigenLayer overview as the canonical anchor, EigenLayer AVS Mechanism · Operator · Slashing · EIGEN Arbitration for the slashing + AVS economics depth, and EigenLayer L1 bootstrapping for how new chains use restaking to launch security. For the validator economics meta-context, see BFT validator economics overview and BFT validator economy four variables. For consensus models that restaking touches, see threshold BFT consensus Rust implementations and DAG-BFT vs Chain-BFT. For the EIP-7251 institutional staking lane that interacts with restaking economics, see Pectra EIP-7251 · institutional staking. For CEX exposure to LRT and the “liquid staking → restaking → CEX leverage” stack, see Liquid Staking + Restaking Ecosystem + CEX Exposure. Cross-link to cross-chain five-pole comparison matrix when reasoning about Hyperlane / LayerZero / Wormhole as AVS consumers.

Why this matrix matters

Restaking in 2023-2024 年 was EigenLayer’s single-player concept of an “ETH crypto-economic security marketplace.” By 2024-2025 年 it evolved into multi-protocol competition: Symbiotic offers multi-asset restaking (not only ETH) · Karak centers developer UX + universal restaking · Mellow is an LRT vault aggregator · EtherFi inherits the LST market through LRT (eETH) · Lido itself also enters through CSM (Community Staking Module) and stVaults.

For developers (AVS builders · for example Hyperlane / EigenDA / Lagrange), which restaking layer to choose determines: operator pool depth · true enforceability of slashing · scale of economic security from TVL backing · governance predictability. For users (stETH holders considering restake), which LRT (eETH · ezETH · pufETH · mevETH, etc.) to choose determines: revenue sharing · slashing risk exposure · withdrawal liquidity · whether it is listed on CEXs.

For regulators (SEC / ESMA / FSA · see Liquid Staking + Restaking Ecosystem + CEX Exposure), whether restaking constitutes a “security” / “investment contract” is an issue awaiting judgment in 2026 年, and protocols have different regulatory exposures because their governance + revenue structures differ.

This matrix compares 6 active restaking infrastructures cross-sectionally · 10 axes · 2026-Q2 data snapshot. The AVS landscape is an independent section · listing the economic image of 5 major AVSs. Stand-alone LRTs (Renzo / Puffer / Kelp / Swell) are not listed independently in the comparison; some are mentioned inside the EtherFi / Mellow analysis.

Per-protocol sections

EigenLayer

Restaking model: Operator-delegated (the staker chooses an operator, and the operator chooses which AVSs to opt into). Supports native ETH restaking (through EigenPod) · LST restaking (8 types of strategies such as stETH · rETH · cbETH) · EIGEN token native restaking.

TVL May 2026: ~$14B (estimate · at the 2025 high reached $20B+ · 2026 saw a mild adjustment from some LRT exits + yield compression).

Slashing implementation: 2025-Q2 mainnet slashing enabled. AVSs define custom slashing conditions (see EigenLayer AVS mechanism). Actual slash events ≤ 5 件 (2026-Q2) · cumulative slash amount < $5M (mainly operator misconfiguration · not malice).

AVS economics: ~40 AVSs active · including EigenDA · Hyperlane · Espresso · Lagrange · Witness Chain · AltLayer · Brevis · etc. Operator rewards come from AVS charges (EIGEN + each AVS native token) · adding annualized +1-3% on top of base ETH staking yield.

Correlation risk: Rehypothecation coefficient 2.5-3x (each 1 unit of ETH opts into an average of 2.5-3 AVSs) · Vitalik has warned multiple times about “system-wide slashing cascade” risk. EIGEN inter-subjective dispute is a mitigation mechanism, but it itself depends on token holder vote (which can be capital-manipulated).

Operator concentration: Top 5 operators account for ~50% of TVL (Coinbase Cloud · Figment · P2P · Stakefish · DSRV) · the remainder is distributed across ~200 operators.

Governance: EigenLabs (team-led) · EIGEN token + Operating Token holders · Security Council 7-of-13. After the 2025 EIGEN voting launch, decentralization proceeds gradually.

Audit: Trail of Bits · ChainSecurity · OpenZeppelin · Sigma Prime · Code4rena · Cantina. Multiple rounds of independent audit before slashing was enabled.

Lido stake overlap: ~25% of EigenLayer TVL comes from the stETH strategy (the rest is mainly native ETH via EigenPod + cbETH · rETH).

Regulatory exposure: The SEC has not made a clear statement on whether the EIGEN token constitutes securities · in 2024 年 EigenLayer actively delayed EIGEN airdrops to US users. Under EU MiCA, restaking yield may be classified as a “tokenized money market service” · awaiting judgment.

Symbiotic

Restaking model: Multi-asset restaking (not only ETH · supports any ERC-20 , including LSTs · stablecoins · project tokens). Modular vault architecture · vault curators custom-define collateral + AVS opt-in strategy.

TVL May 2026: ~$3.2B (2024-09 alpha launch · 2025 fast growth · second-largest restaking in 2026-Q1 · some TVL is non-ETH assets, so direct comparison with EigenLayer needs adjustment).

Slashing implementation: Supports vault-level slashing from day 1 (2025-Q1 mainnet enabled). Operator slashed = the entire vault collateral is damaged · but vault curators can customize the slashing strategy (ratio · trigger conditions).

AVS economics: ~15 AVSs active (2026-Q2) · main pushes are Ethena USDe / sUSDe restake · 1 of Hyperlane ISM · Mellow vault deep integration · GearBox credit account, etc. Operator rewards are diversified (stablecoin · vault token · AVS native).

Correlation risk: Theoretically higher than EigenLayer (multi-asset classes · complex slashing trigger conditions) · but in practice, diversified assets can diversify single-ETH risk. Vault curator selection is the risk anchor.

Operator concentration: Top 5 operators ~60% (P2P · Kiln · Chorus One · Everstake · DSRV) · concentration is higher than EigenLayer.

Governance: Symbiotic Labs team-led · token + DAO governance planned in 2026-Q4 . Vault curators are de facto sub-governance (each vault is independent).

Audit: OpenZeppelin · Sigma Prime · Cantina contest · early stage · audit coverage is less than EigenLayer but continues to expand.

Lido stake overlap: ~40% of TVL comes from LST vaults (stETH · rETH · wstETH · etc.). Mellow, as Symbiotic’s main vault curator, routes LRT funds into Symbiotic AVSs.

Regulatory exposure: Multi-asset structure makes Symbiotic’s regulatory positioning more complex (some vaults include stablecoins = closer to money market regulation) · but the impact of single-asset regulatory events is also more limited.

Karak

Restaking model: Universal restaking (supports ETH · BTC · USDC · K2 is Karak’s own L2) · emphasizes developer UX (“any asset can secure any service”).

TVL May 2026: ~$1.0B (2024 launch · 2025 growth but slower than Symbiotic · 2026 rises with BTC restaking module + K2 integration).

Slashing implementation: Enabled in 2025-Q3 · but actual AVS count is low · triggers slash events ≤ 2 件.

AVS economics: ~10 AVSs active · mainly services incubated by the Karak team itself (K2 L2 sequencer / DA) · few third-party AVSs.

Correlation risk: Correlation is complex under the multi-asset model · but TVL scale is small · systemic risk is low for now.

Operator concentration: Top 5 operators ~70% (Karak self-operated + a small number of partners).

Governance: Karak Foundation-led · KAR token planned to launch in 2026-Q3-Q4 .

Audit: Sigma Prime · Halborn · Cantina (partial).

Lido stake overlap: ~20% of TVL comes from stETH (Karak has a larger share of BTC + stablecoin restaking).

Regulatory exposure: BTC restaking triggers SEC interest in “BTC yield products.” USDC restaking is closer to money market regulation. Overall regulatory surface is broader than EigenLayer.

Mellow

Restaking model: LRT vault aggregator (not base-layer restaking, but a vault layer built on top of EigenLayer + Symbiotic) · Mellow vault users deposit stETH / wstETH · the vault automatically restakes into EigenLayer/Symbiotic AVSs · revenue sharing + slashing risk are pooled.

TVL May 2026: ~$1.8B (2024-Q4 launch · 2025-2026 fast growth · Symbiotic’s largest vault curator · also deploys some EigenLayer vaults).

Slashing implementation: Slashing is socialized at the vault layer (a slash of 1 AVS is borne by all vault holders according to shares) · users implicitly accept portfolio slashing risk.

AVS economics: Mellow vaults select multiple AVS pools to optimize risk-adjusted yield · users receive a basket of “mixed AVS rewards.”

Correlation risk: Highest layer: vaults hold multiple AVSs at the same time · slash cascade risk inside a single vault is pronounced. Mellow mitigates through vault diversification (different AVS pools in different vaults).

Operator concentration: Borrows the EigenLayer / Symbiotic operator set · does not introduce new concentration itself.

Governance: Mellow DAO + MLW token (2025-Q2 launch) · vault curator team as sub-governance.

Audit: Sigma Prime · Spearbit · Code4rena · joint audit with Symbiotic.

Lido stake overlap: ~100% (Mellow vaults are predominantly based on stETH, wstETH, and other LST sets) · the largest single route from Lido → restaking.

Regulatory exposure: Vault structure is close to a “managed fund” · may be classified as a collective investment scheme under EU MiCA / SEC. Mellow actively avoids US users.

EtherFi LRT (eETH / weETH)

Restaking model: Integrated LST + restaking: users deposit ETH and receive eETH (rebasing) or weETH (wrapped non-rebasing) · EtherFi simultaneously stakes the underlying ETH (beacon chain validator)+ restakes it (EigenLayer + Symbiotic + Karak partial). Leader in the LRT market.

TVL May 2026: ~$5.5B (2024 high $7B · 2026 mild adjustment) · largest LRT.

Slashing implementation: EtherFi operates nodes itself (node operator network · strictly screened) · slash risk is borne by the EtherFi insurance fund + protocol reserve. Users face it through eETH price reflecting slash losses.

AVS economics: EtherFi chooses to restake ETH into mainstream AVSs (EigenDA · Hyperlane · Espresso · Lagrange) · returns are reflected in eETH yield (annualized ~4-6% · including base staking 3-4% + restaking premium 1-2%).

Correlation risk: eETH holders are naturally exposed to the 3 -layer slash cascade of EigenLayer + Symbiotic + Karak. EtherFi insurance fund of ~$50M acts as first-loss buffer.

Operator concentration: EtherFi node network ~200 operators · some directly operated by EtherFi · some outsourced (Pier Two · ChainSafe · Validation Cloud, etc.).

Governance: EtherFi DAO + ETHFI token (2024-Q2 launch) · Foundation + Security Council 7-of-15.

Audit: Spearbit · Halborn · Certora · Hats Finance bug bounty.

Lido stake overlap: EtherFi is a direct competitor to Lido (LST market share) · EtherFi does not use stETH (it operates its own validators) · but when some EtherFi stake is put into EigenLayer, it sits inside the same AVS pools as stETH.

Regulatory exposure: LRT is a key SEC focus (2024 SEC chair Gensler publicly mentioned “staking + restaking = potential security”). EtherFi actively geo-fences US retail users.

Lido CSM (Community Staking Module)

Restaking model: Lido itself did not enter restaking for a long time (rejected multiple times in Lido v2 governance) · but announced CSM (Community Staking Module) in 2024 年, allowing small operators to join the Lido validator set · while also exploring stVaults (2026 plan) so stETH holders can opt into restaking. Lido v3 (2026 roadmap) formally introduces a restaking hook.

TVL May 2026: ~$280M (CSM early adoption · rapid growth expected after 2026 Q3-Q4 stVaults launch). Total Lido stETH TVL ~$32B (restaking portion < 1%).

Slashing implementation: CSM operators bring their own bond (2-4 ETH) as collateral · slash first deducts the bond · any shortfall is borne by the Lido insurance fund. stVaults restaking slash is borne by vault users according to shares.

AVS economics: Lido stVaults plan to support mainstream AVSs on EigenLayer + Symbiotic · but Lido DAO strictly manages the AVS whitelist (only fully audited + governance-stable services can connect).

Correlation risk: Lido’s design philosophy is “minimize correlation” · after stVaults opt-in, users implicitly bear it · but Lido DAO limits through AVS whitelist + floating caps.

Operator concentration: Lido main set ~30 operators (institutions)+ CSM ~200 small operators. CSM improves decentralization · introduces long-tail home stakers.

Governance: Lido DAO + LDO token + public referendum · highest governance complexity (connecting any AVS to stVaults requires passing an LDO vote).

Audit: Sigma Prime · ChainSecurity · OpenZeppelin · Certora · Statemind · MixBytes · audit coverage is the deepest in the industry.

Lido stake overlap: Lido is the stETH issuer · “Lido stake overlap” here = 100%. Secondary restaking of stETH in EigenLayer / Symbiotic / Mellow / EtherFi accounts for 25-40% of the whole restaking ecosystem, and Lido is effectively the source of restaking liquidity (see Liquid Staking + Restaking Ecosystem + CEX Exposure).

Regulatory exposure: Lido DAO governance structure is close to a “decentralized issuer” from the SEC perspective · but LDO token holder voting + governance revenue distribution may constitute securities. Whether stETH constitutes an “asset-referenced token” under EU MiCA is awaiting judgment.

AVS landscape — economic image of 5 major AVSs

AVSTypeMain backing2026 revenue estimateTVL backingCustomers
EigenDAData AvailabilityEigenLayer~$120M ARR~$6B ETHMantle · Movement · Cyber · Rivalz · multiple ZK Stack hyperchains
HyperlaneCross-chain messaging ISMEigenLayer + Symbiotic + own~$30M ARR (messaging fee distribution)~$1.5B ETH (EigenLayer ISM) + some SymbioticEclipse · Celestia · Renzo · multiple app-chains
EspressoShared sequencerEigenLayer~$15M ARR (early)~$800M ETHPolygon zkEVM · zkSync ZK Stack · Caldera · AltLayer rollup-as-a-service
LagrangeCoprocessor + state proofEigenLayer~$8M ARR~$500M ETHLayerZero · Polyhedra · multiple DeFi protocols
Witness ChainPOI (Proof of Inference)/ DePIN verificationEigenLayer~$5M ARR (experimental)~$200M ETHDePIN projects + AI inference verification

2 types of AVS economic model:

  • Direct fee: users / customers are charged (EigenDA customers pay DA fees · Hyperlane customers pay messaging fees) → distributed to operators
  • Token incentive: AVS has its own token · provides initial emission to operators as bootstrapping (Espresso · Lagrange · Witness Chain)

Rehypothecation impact: The top 3 AVSs (EigenDA · Hyperlane · Espresso) all draw backing mainly from EigenLayer · it is normal for each 1 unit of ETH to secure 3 AVSs simultaneously. The “cascade risk” Vitalik warns about mainly refers to this situation.

Big comparison matrix table

6 restaking infrastructures × 10 axis comparison (state as of 2026-Q2 ):

ProtocolRestaking ModelTVL May 2026SlashingAVS EconomicsCorrelation RiskOperator ConcentrationGovernanceAuditLido OverlapRegulatory
EigenLayerOperator-delegated · native ETH+LST+EIGEN~$14B2025-Q2 enabled · ≤5 events · <$5M cumulative~40 AVS · EIGEN+AVS token rewards · +1-3% on baseRehypothecation 2.5-3x · cascade concernTop 5 ~50%EigenLabs+EIGEN+SC 7/13Trail of Bits · ChainSecurity · OZ · Sigma Prime · C4 · Cantina~25% stETHSEC EIGEN security not stated · EU MiCA awaiting judgment
SymbioticMulti-asset (ETH+LST+stablecoin+token) · modular vault~$3.2BDay 1 design · 2025-Q1 enabled · vault-level~15 AVS · stablecoin/vault/AVS token rewardsComplex multi-asset · vault curator anchorTop 5 ~60%Symbiotic Labs · 2026 Q4 token+DAOOZ · Sigma Prime · Cantina · early stage~40% stETH (Mellow vault main route)Complex multi-asset regulatory surface
KarakUniversal (ETH+BTC+USDC) · K2 L2 integration~$1.0B2025-Q3 enabled · ≤2 events~10 AVS · Karak-incubated centeredMulti-asset complex · small TVL, low risk for nowTop 5 ~70% (high concentration)Karak Foundation · KAR 2026 Q3-Q4Sigma Prime · Halborn · Cantina partial~20% stETHBTC restaking SEC interest · USDC close to MM regulation
MellowLRT vault aggregator (on EigenLayer+Symbiotic) · socialize slashing~$1.8BVault-level socializeMulti-AVS pool basket rewardsHighest (in-vault cascade)Borrows underlying · no new concentrationMellow DAO+MLW 2025-Q2 · vault curator sub-govSigma Prime · Spearbit · C4 · joint with Symbiotic~100% (predominantly LST-based)Close to managed fund · avoids US
EtherFi LRTIntegrated LST+restaking · own nodes+multi-layer restake~$5.5B (largest LRT)EtherFi node responsibility · insurance fund first-lossEigenDA+Hyperlane+Espresso+Lagrange · annualized 4-6%3 layer slash cascade · insurance ~$50M bufferEtherFi network ~200 operatorEtherFi DAO+ETHFI 2024 Q2 · SC 7/15Spearbit · Halborn · Certora · Hats bountyLido competitor · own validators · indirect AVS pool overlapLRT is SEC focus · US retail geo-fence
Lido CSM / stVaultsCSM home staker + stVaults restaking hook (2026 plan)~$280M CSM · total stETH $32BCSM operator bond first · insurance backup · stVaults users bearStrict AVS whitelist (DAO vote)Lido philosophy minimize · AVS cap limitsCSM ~200 small operators + main set 30Lido DAO+LDO+referendum (heaviest governance)Sigma Prime · ChainSecurity · OZ · Certora · Statemind · MixBytes · industry deepest100% (Lido itself) · stETH is restaking liquidity sourceLDO security awaiting judgment · stETH MiCA ART awaiting judgment

How to read the matrix:

  • Restaking model 3 quadrant comparison: operator-delegated (EigenLayer · Karak · Symbiotic) → direct delegation · vault aggregator (Mellow) → wraps other restaking · integrated LST+restaking (EtherFi · Lido stVaults) → users receive LRT tokens
  • TVL vs operator concentration tradeoff: EigenLayer has the largest TVL · highest operator dispersion · Karak has small TVL · high concentration · Mellow borrows others’ operators and adds no new concentration
  • Lido overlap is a hidden systemic risk: secondary restaking of stETH in EigenLayer / Symbiotic / Mellow / EtherFi · if stETH price depegs or a Lido protocol incident occurs, the whole restaking ecosystem is affected simultaneously
  • Actual state of slashing: even after 1+ years from enablement, actual slash events ≤ 10 件 · cumulative amount < $10M · far below the “$100M+ cascade” scenario Vitalik warned about · but the sample size is small, so one cannot conclude that risk has been priced

Boundary cases / future trajectory

**Interaction between restaking and EIP-7251 ** (see Pectra EIP-7251 · institutional staking):

  • EIP-7251 expands validator max effective balance from 32 → 2048 ETH, enabling large institutions to consolidate validators and lower operating cost.
  • Institutional validators can simultaneously restake large amounts of ETH · further raising operator concentration. EigenLayer has already confirmed in 2025 年 a small rise in the top 5 operator ratio.
  • Over the long term, if institutional staking (SWIFT custodians · BNY · Coinbase Custody · BlackRock · see Liquid Staking + Restaking Ecosystem + CEX Exposure) enters restaking, TVL may rise from 2026 年’s $20B class to, in 2027-2028 年, $50B+.

Lido v3 restaking connection: Lido DAO 2026-2027 stVaults mainline development · once connected, stETH becomes a direct restaking primitive and secondary wrappers are no longer necessary. This would mean:

  • Some value of Mellow / EtherFi is absorbed (stETH holders directly opt in · intermediary vault unnecessary)
  • But Lido’s strict AVS whitelist also limits bootstrapping of long-tail AVSs
  • In the long term, Lido + EigenLayer form a de facto duopoly, while Symbiotic / Karak / Mellow / EtherFi handle peripheral subdivisions

Evolution of AVS business models:

  • 2024-2025 stage: Most AVSs depend on token incentive bootstrapping (distributing their own tokens to operators) · real fee revenue is small.
  • 2026 stage: Head AVSs (EigenDA · Hyperlane · Espresso) begin to have real fee revenue, but AVS-to-operator distribution ratios do not match (EigenDA gives 90%+ to operators · Hyperlane is 50/50 · Espresso is still bootstrapping).
  • 2027+ trend: The AVS market may split into “DA commoditization” (EigenDA / Celestia / Avail price war) and “differentiated services” (Espresso shared sequencer · Lagrange coprocessor · Witness Chain DePIN verification).

Rehypothecation systemic risk:

  • Vitalik in 2024 年 repeatedly warned against overextending EigenLayer into “application-layer consensus”: concern about slashing cascade when each 1 unit of ETH simultaneously secures 5+ AVSs.
  • Actual data: 2026-Q2 average rehypothecation coefficient 2.5-3x · far below the theoretical maximum (any number of AVSs) · because most ETH opts into only 2-3 head AVSs.
  • But LRT vaults such as EtherFi / Mellow opt into multiple AVSs by default · users implicitly bear cascade. Mellow vault internal diversification mitigates · but cross-vault correlation still exists (the same operator runs multiple AVSs at the same time).

Symbiotic multi-asset strategy vs EigenLayer ETH-pure:

  • EigenLayer maintains ETH (+LST) purity · because Ethereum community consensus + ETH is the deepest crypto-economic security pool.
  • Symbiotic accepts any asset (including stablecoins · project tokens), letting AVSs self-select risk profiles.
  • Over the long term, Symbiotic may win customers EigenLayer cannot get in “non-ETH-native AVSs” (BTC cross-chain · DePIN · AI inference); EigenLayer keeps its advantage in “ETH-aligned AVSs” (customers are ETH-native dApps).

Slashing case studies being collected:

  • 2025-Q3 EigenLayer operator (mid-sized node) 1 件 triggered a ~$200K slash through validator-key misconfiguration · operator compensated at its own cost.
  • 2025-Q4 EigenDA 1 件 quorum-shortage event triggered a ~$80K collective slash · the AVS mitigated itself.
  • 2026-Q1 Symbiotic 1 件 vault configuration bug caused a ~$1M slash · vault curator compensated with insurance.
  • Actual slash cases cumulative < $5M · but each case raised community vigilance · promoted stronger audits + insurance pools.

Regulatory boundary:

  • The SEC in 2024 年 repeatedly stated that “staking-as-a-service” is close to securities (already took action against Coinbase · Kraken) · restaking is a superset of staking · same risk.
  • Under EU MiCA, LST may be classified as ART (asset-referenced token) · LRT is more complex (base layer is LST · upper layer adds restaking yield) · regulatory route is unclear.
  • Japan FSA 2025 public statement: staking-as-a-service is an entry subject for “ERS (Electronic Recordable Securities)” · restaking should also follow this.
  • Before institutional customers (BNY · BlackRock · J.P. Morgan) enter restaking, regulatory clarity is needed · timeline is at least 2027+.

Formal verification / engineering quality:

  • EigenLayer has the most complete audit coverage in restaking (6+ independent audit firms + multiple rounds before slashing enablement)
  • Lido CSM / stVaults have the industry’s highest audit depth (7+ firms + Certora formal verification)
  • Symbiotic / Karak / Mellow / EtherFi are still catching up in audit coverage · some early audits do not include slashing logic

Relation between restaking and DAG-BFT / threshold consensus: Restaking lets AVSs launch without newly recruiting validators · but AVSs’ own consensus-algorithm choice remains important (see DAG-BFT vs Chain-BFT and threshold BFT consensus Rust implementations). Espresso shared sequencer uses HotShot (a DAG-BFT variant) · Lagrange coprocessor leans toward a threshold-signature model. AVS consensus choice affects operator real workload + revenue structure, and in reverse affects which restaking layer is welcomed by operators.

Intersection of validator economy 4 variables: See BFT validator economy four variables: the 4 variables of security budget · stake mobility · validator count · slashing severity are all rewritten at the restaking layer. Restaking expands stake mobility from a single chain to multi-AVS · reuses validator count instead of newly recruiting · but complicates slashing severity (the same ETH has different slash conditions across multiple AVSs). This is why Vitalik repeatedly warns that restaking should not be overextended into “application-layer consensus”: when the number of AVSs exceeds operator monitoring bandwidth, validator economic variables are no longer well-defined.

Sources