Skip to content
Decision guide // updated July 2026

EigenLayer vs Symbiotic vs Babylon: Choosing a Restaking Layer

We build AVSs and restaking integrations day to day, so here is what actually separates EigenLayer, Symbiotic, and Babylon as shared-security layers in 2026.

TL;DR

Your collateral decides most of this: ETH-based services go to EigenLayer or Symbiotic, native BTC to Babylon. EigenLayer (now EigenCloud) leads by a wide margin on ecosystem, operators, and liquidity, with live slashing and redistribution, but sits mostly on ETH and carries concentration risk. Symbiotic is the permissionless, collateral-agnostic number two: any ERC-20, isolated per-network vaults, full config control, though it has no token yet and a smaller ecosystem. Babylon is the Bitcoin-native alternative, letting native BTC secure networks without wrapping or bridging. Karak, once a third contender, has effectively left restaking (rebranded to OpenGDP), a reminder to weigh vendor viability. And the first question is whether you need restaking at all.

Protofire builds actively validated services (AVSs) and restaking integrations on EigenLayer as part of our services, so we have a commercial stake in this comparison. We are not affiliated with Symbiotic, Babylon, or Karak, and each section notes where an alternative beats EigenLayer. Figures come from the protocols' own docs and DefiLlama, pulled in July 2026; restaking TVL moves fast, so treat them as a dated snapshot. See our other decision guides, including rollup stacks and cross-chain messaging.

Scorecard

At a glance

01EigenLayer02Symbiotic03Babylon
CollateralETH + ETH LSTs (+ ERC-20s via operator sets)Any ERC-20 (isolated vaults)Native BTC (self-custodial)
Security modelPooled restaking, opt-in operator setsPer-network isolated vaultsBitcoin staking + finality providers
Mainnet slashingLive (Apr 2025) + redistributionLive at launch (Jan 2025)Finality-provider slashing (BTC-enforced)
Native tokenEIGEN (live)None (points only)BABY (live)
Live TVL (DefiLlama, Jul 2026)~$5.0B~$329M (base-layer; ~$550M self-reported)~$3.3B (BTC-denominated)
OnboardingPermissionless via operator setsFully permissionless, modularVia Babylon Genesis / BSNs
Best atLargest AVS ecosystem & liquidity (ETH)Flexible collateral, full config control (ETH)Bitcoin-denominated security
Benchmark

Secured value, mid-2026

EigenLayer~$5.0B
Babylon (BTC)~$3.3B
Symbiotic~$329M
Karak~$6.6M

Approximate secured value (base-layer TVL) as of July 2026, per DefiLlama, on a log scale (bars are relative, not linear). Base-layer TVL excludes ETH routed through liquid restaking tokens, which DefiLlama counts as separate protocols, so "total secured" is higher; Babylon's figure is BTC-denominated and moves with the Bitcoin price. Karak is shown to illustrate its exit from restaking, not as a live option.

In detail

The three shared-security layers

01

EigenLayer (EigenCloud)

The market leader, now a verifiable-cloud platform
Strengths
  • +The largest ecosystem: the most AVSs, operators, and restaked liquidity, with broad LRT support
  • +Slashing and redistributable slashing are both live on mainnet, with opt-in operator sets that isolate each service's risk
  • +EigenDA is a mature, production data-availability product (100 MB/s), and EIGEN's intersubjective forking backs faults that can't be proven on-chain
Trade-offs
  • Concentrated on ETH and ETH LSTs, and historically the large majority of all Ethereum restaking TVL, a systemic-concentration and re-hypothecation concern
  • Newer EigenCloud products (EigenCompute, EigenAI) are mainnet alpha, not GA; EigenVerify is not yet confirmed on mainnet
  • TVL fell from a roughly $20B 2024 peak to about $5B; much of the early total was points farming, not durable demand

EigenLayer rebranded to EigenCloud in June 2025 and now markets itself as a "verifiable cloud": EigenLayer is the restaking security layer, and EigenDA, EigenCompute, and EigenAI are the products built on top. Slashing went live in April 2025 using operator sets and unique stake (an operator allocates a specific slice of stake to one service's set, so only that service can slash it), and redistributable slashing followed in July 2025, letting slashed ERC-20s, LSTs, or USDC be redirected rather than only burned, which is what makes lending- and insurance-style AVSs possible (native ETH and EIGEN are not yet redistribution-eligible).

EigenDA V2 runs at 100 MB/s on mainnet and is the dominant AVS by secured value, with LayerZero and Infura's DIN among the launch slashing partners. The caveats: TVL is roughly $5B as of mid-2026, down from a ~$20B 2024 peak as points farming ended and slashing repriced risk; EigenLayer still holds the large majority of Ethereum restaking TVL, which concentrates Ethereum's economic security and revives the re-hypothecation critique; and the newer EigenCloud services (EigenCompute, EigenAI) are mainnet alpha, with EigenVerify not yet confirmed live. It is the default when you want the largest operator set and liquidity on Ethereum.

02

Symbiotic

Permissionless, collateral-agnostic shared security
Strengths
  • +Accepts any ERC-20 as collateral, not just ETH, in isolated per-network vaults with network-defined slashing and rewards
  • +Fully permissionless and modular: spin up a vault or network with no whitelist, and the core contracts are immutable
  • +Live slashing since its January 2025 mainnet, with early adopters including Ethena and Hyperlane
Trade-offs
  • No token as of mid-2026 (a two-year points program with no announced tokenomics), so incentives carry uncertainty
  • A much smaller ecosystem than EigenLayer (~$329M vs ~$5B), and arbitrary-ERC-20 collateral is only as sound as that asset's liquidity
  • In July 2026 it broadened from restaking into collateral markets (Core V2), adding surface area that is still very new

Symbiotic launched a feature-complete mainnet with live slashing in January 2025, and its differentiator is flexibility: any ERC-20 can be collateral, held in isolated vaults, with operators opting into networks (its term for AVSs) and resolvers acting as a slashing veto. The core contracts are immutable, and onboarding is fully permissionless. Symbiotic Relay lets a network derive its validator set epoch by epoch with BLS aggregation and optional ZK compression, so verification cost stays flat as the set grows. Ethena (which restaked ENA and sUSDe as the first non-ETH collateral) and Hyperlane are among the early adopters, with Mellow curating vaults on top; Symbiotic reports securing more than $550M across applications (DefiLlama's base-layer count is lower, around $329M).

Two things to weigh. First, there is still no Symbiotic token as of mid-2026, only a points program, so anyone counting on a liquid security token or emissions is working from an implied promise. Second, on July 1, 2026 Symbiotic launched Core V2 and repositioned from restaking to collateral markets, routing idle vault capital into venues like Aave and Morpho and targeting insurance, credit, and tokenized assets. That is a genuine broadening of scope, and it is weeks old. Symbiotic is the pick when you need collateral flexibility, isolated risk per network, and full control of parameters, and can accept a smaller, younger ecosystem.

03

Babylon

Bitcoin-native staking and shared security
Strengths
  • +Lets native BTC secure networks with no wrapping, bridging, or third-party custody of the coins
  • +Second-largest security marketplace by TVL (~$3.3B), and a live BABY token for gas, governance, and dual staking
  • +Exports Bitcoin-backed finality to other chains via Bitcoin Supercharged Networks (BSNs)
Trade-offs
  • A different domain: it secures Babylon Genesis and BSN chains, not the Ethereum AVS marketplace, so consumers differ
  • BTC-denominated TVL swings with the Bitcoin price (down from a ~$5.6B 2026 peak)
  • Finality-provider slashing and the BSN model are newer and less battle-tested than a mature validator set

Babylon is the Bitcoin-native answer to shared security. Rather than restaking ETH, it lets holders stake native BTC through Bitcoin scripts, self-custodially, to provide slashable economic security, with finality providers whose double-signing can be penalized by Bitcoin-enforced slashing. It secures Babylon Genesis, a Cosmos-SDK Layer 1, and exports that Bitcoin-backed finality to other chains through Bitcoin Supercharged Networks, using a dual-staking model that pairs BTC with its live BABY token.

By secured value it is around $3.3B, larger than Symbiotic and second only to EigenLayer, though because the collateral is BTC the dollar figure moves with the Bitcoin price (it was reported near $5.6B earlier in 2026). The trade-off is domain: Babylon is not a drop-in substitute for an Ethereum AVS. If your product, users, or the security you want to buy live in the Bitcoin economy, or you specifically want BTC rather than ETH backing your network, Babylon is the natural choice; if you need the Ethereum AVS ecosystem and its operators, EigenLayer or Symbiotic fit better.

Verdict

Which should you use?

If your priority is
The largest ecosystem, operators, and liquidity on Ethereum

you want the deepest AVS marketplace and LRT liquidity plus live redistribution, and can accept ETH concentration and that the newer EigenCloud products are still alpha.

EigenLayer
If your priority is
Flexible collateral and isolated, self-configured security

you want any-ERC-20 collateral, per-network vaults, and full parameter control, and can accept a smaller ecosystem and no token yet.

Symbiotic
If your priority is
Bitcoin-denominated security

your collateral or users are BTC-native and you want Bitcoin, not ETH, backing your network, without wrapping or bridging.

Babylon
If your priority is
You may not need restaking at all

you already run a healthy validator set, or your service doesn't need externally attributable, slashable stake, so a shared-security dependency would add risk without buying much.

Own validator set
Also consider

The wider restaking map

  • Karak (now OpenGDP): once discussed as the third major restaking layer, it has effectively left the category. karak.network now redirects to opengdp.network, the project rebranded away from restaking, and its TVL collapsed to roughly $6.6M from 2024 peaks above $700M. Treat it as a cautionary example of vendor viability, not a live option.
  • Jito (Solana): if you build on Solana, Jito Restaking is the native option, but it is still very small (~$14M), and note that Jito's headline figure is JitoSOL liquid staking, a much larger and separate product, not restaking.
  • LRTs and the wrapper layer (Ether.fi, Kelp, Renzo, Puffer): most restaked ETH reaches the base layers through liquid restaking tokens, and most catastrophic recent losses happened here, not in the base protocols, most notably Kelp's ~$292M rsETH bridge exploit in April 2026. If you consume restaked security through an LRT, you inherit its curator and bridge risk on top of the base layer.

FAQ

What is restaking?
Restaking lets assets already staked to secure one network, typically ETH staked on Ethereum through a liquid staking token, be re-committed to secure additional services at the same time, for extra rewards and extra slashing risk. Those additional services (AVSs, or actively validated services, on EigenLayer; networks on Symbiotic) are things like oracles, bridges, data-availability layers, and coprocessors that need their own economic security but do not want to launch a token and validator set from scratch. Babylon applies the same idea to native Bitcoin. The core trade-off is that one pool of collateral now backs several obligations at once.
What is the difference between EigenLayer and Symbiotic?
Both let a new network buy economic security from restakers instead of bootstrapping its own validator set, but the models differ. EigenLayer centers on ETH and ETH LSTs (with more assets via operator sets), pools security with opt-in operator sets, has a live EIGEN token, and by mid-2026 holds the large majority of Ethereum restaking TVL (~$5B) with the biggest operator and AVS ecosystem. Symbiotic accepts any ERC-20 as collateral in isolated per-network vaults with network-defined slashing, is fully permissionless with immutable core contracts, and has no token yet (~$329M by DefiLlama's base-layer count). In short: EigenLayer for ecosystem depth and liquidity on ETH; Symbiotic for collateral flexibility, risk isolation, and configurability.
Is EigenLayer still the biggest restaking protocol in 2026?
Yes. Roughly $5B in base-layer TVL versus about $329M for Symbiotic means it holds most Ethereum restaking TVL. But two things are true at once: it still leads, and the whole sector deflated from its 2024 peak (EigenLayer was near $20B) as points farming ended and live slashing repriced risk. It also rebranded to EigenCloud in 2025, so "EigenLayer" now refers to the restaking layer inside a broader verifiable-cloud platform.
What happened to Karak?
Karak, built by Andalusia Labs and once counted among the top restaking layers after a $48M Series A at a $1B+ valuation in 2023, has effectively exited restaking. Its domain now redirects to OpenGDP, a rebrand toward real-world-economy infrastructure with no restaking, K2, or DSS messaging, and its TVL collapsed to about $6.6M from 2024 peaks above $700M, with no live token and no confirmed mainnet slashing. The practical lesson for anyone choosing a security layer: the team's continued commitment to the product is itself a risk axis.
Do I actually need restaking or an AVS?
Often not. Restaking exists to give a new service, an oracle, bridge, data-availability layer, coprocessor, or fast-finality/pre-confirmation network, externally attributable and slashable economic security without launching its own token and validator set. If you already run a healthy validator set, or your service doesn't need slashable outside stake, a shared-security dependency adds trust assumptions and smart-contract risk without buying much. Start by deciding whether your product genuinely needs bought security; only then does the EigenLayer-vs-Symbiotic-vs-Babylon question matter.
Is restaking safe?
It carries specific risks. At the base layer, restaked collateral secures many services at once, which revives a re-hypothecation concern (the same ETH backing multiple liabilities, with cascading-liquidation risk) and, given EigenLayer's dominance, concentrates Ethereum's economic security. Slashing is now live on EigenLayer and Symbiotic, but each service defines its own conditions, so risk is non-uniform and must be checked per network. In practice, most of the largest recent losses have been in the LRT and bridge wrapper layer, not the base protocols, for example Kelp's ~$292M rsETH bridge exploit in April 2026. Building on restaking means designing for slashing conditions, collateral quality, and wrapper risk deliberately.

Reviewed by Luis Medeiros, Field CTO at Protofire. Last updated: July 2026.

Build it

We build actively validated services and restaking integrations, including the operator set, slashing conditions, collateral design, and the off-chain infrastructure that runs them.

EigenLayer AVS development

Book a call with Alejandro Losa

Building an AVS or a restaking integration? Talk to us about the right shared-security layer and the slashing, operator, and collateral design around it.

Protofire 2026. All rights reserved

Message us on Telegram