This ranking covers restaking base layers — the protocols that let staked capital secure additional networks and that decide who gets slashed when something goes wrong. It is not a ranking of liquid restaking tokens, which are the wrappers built on top of these layers and have their own page. The category matured sharply through 2026: slashing moved from roadmap to production on the leaders, which turned restaking from a yield story into a risk-pricing one. We ranked 7 protocols on restaked capital, live network demand, slashing design, collateral breadth and decentralisation, using magnitude bands rather than point figures.
We scored each restaking protocol across 8 weighted criteria: restaked capital depth (15%), live network demand measured by networks actually paying for security rather than announced (20%), slashing design and whether it is in production (20%), collateral breadth and quality (10%), operator set size and decentralisation (10%), token economics and fee capture (10%), integration surface for networks and operators (10%) and 2026 roadmap credibility (5%).
Data sources: DefiLlama and protocol dashboards for restaked TVL, protocol documentation and onchain registries for live network and operator counts, published slashing specifications and audit reports, and our own reading of shipped-versus-announced upgrades. We exclude liquid restaking token issuers, which are ranked separately, and protocols whose slashing remains entirely unspecified, because an unslashable restaking layer is not providing economic security in any meaningful sense.
Critical context: restaking in 2026 is a risk product, not a yield product, and the ranking reflects that. The question that separates these protocols is no longer how much capital they hold but what happens when a network they secure misbehaves — who decides, how fast, and whether the depositor had any way to price that risk in advance. Protocols with production slashing score higher here than protocols with more TVL and slashing still in specification.
Scoring is 0-10 per criterion with a weighted average producing the final score. Score range in this ranking: 6.6 to 8.9. We rank 7 rather than 10 because the base-layer category has only seven credible entries; the obvious candidates to pad the list with are liquid restaking token issuers, and those belong on our LRT ranking instead. Restaking carries slashing risk on top of staking risk, and nothing in this ranking is a recommendation to restake capital you cannot afford to see penalised.
Data as of 2 October 2026. Restaked TVL, live network counts and operator set sizes move continuously, so figures here are stated as magnitude bands rather than quotes — treat them as indicators, not live data. Slashing status is stated as of this review date and is the field most likely to have changed since.
| Criterion | Weight | What we measure |
|---|---|---|
| Live network demand | 20% | Networks actually paying for security, not announced |
| Slashing design | 20% | Whether slashing is in production and how it is governed |
| Restaked capital depth | 15% | Restaked TVL as a measure of available security |
| Collateral breadth | 10% | Assets accepted and their quality as collateral |
| Operator decentralisation | 10% | Operator set size and concentration |
| Token economics | 10% | Fee capture and reward distribution design |
| Integration surface | 10% | Tooling for networks and operators to integrate |
| 2026 roadmap credibility | 5% | Shipped-versus-announced record |
Detailed evaluation for each protocol. Top scores get gold, silver and bronze badges. Scoring details in the methodology section above.
EigenLayer created this category and still holds the most restaked capital and the largest set of networks actually consuming it. The core idea is unchanged and still the clearest: ETH that is already staked can be committed to secure additional services, so a new network rents economic security instead of bootstrapping a token and a validator set from zero. What changed in 2026 is that slashing went live, and that matters more than any TVL figure — before slashing, restaking was a rewards programme; after it, the security is real and so is the downside. The AVS set spans data availability, oracles, coprocessors and bridges, and EigenDA alone gave the protocol a flagship consumer with genuine throughput. The honest weakness is concentration and complexity: operator share is uneven, a depositor delegating to an operator inherits slashing exposure across every service that operator opted into, and reasoning about that aggregate risk is beyond most users. For networks that need security today, it is the deepest market available.
Symbiotic's bet is that the network buying security should choose its own parameters rather than accept a shared framework, and the architecture follows that through further than any competitor. Any ERC-20 can be collateral, each network defines its own slashing conditions, resolver and operator requirements, and vaults sit between depositors and networks as the place where risk is actually configured. For a team with an unusual security requirement — a non-ETH collateral base, a custom resolution path, a specific operator whitelist — this is the only protocol here that accommodates it without a fork. Adoption has been strong among newer networks precisely for that reason, and the Mellow vault ecosystem grew up around it. The honest weakness is the flip side of the same design: permissionless collateral means some vaults are secured by assets whose own liquidity would evaporate in the stress scenario that triggers slashing, and the burden of assessing that falls on the depositor rather than on a protocol-level standard.
Babylon is the only protocol here that put Bitcoin to work as staking collateral without moving it. BTC is locked in a self-custodial timelocked script on Bitcoin itself, and slashing is enforced through an extractable-one-time-signature construction that reveals the staker's key if they double-sign — so the penalty is cryptographic rather than dependent on a bridge, a multisig or a wrapped representation. That matters because every previous attempt to use BTC as security required trusting a custodian or a bridge, and those are precisely the components that have failed most often in this industry. The result is the largest pool of non-ETH restaking collateral in existence, and it gave Cosmos-ecosystem chains a credible alternative to bootstrapping their own token security. The honest weakness is Bitcoin's own constraints: unbonding is slow because it is governed by Bitcoin timelocks, the programmability ceiling is low compared with EVM-based designs, and the set of networks able to consume BTC security is narrower than EigenLayer's.
Karak competes on collateral breadth and on giving networks somewhere to test before they are live. It accepts the widest range of assets of any protocol here — LSTs, LRTs, stablecoins and selected ERC-20s — which lowers the barrier for a depositor who holds something other than ETH and does not want to swap into it. K2, its own L2, functions as a staging environment where a distributed secure service can run against real restaked capital before committing to mainnet parameters, and that is a genuinely useful piece of infrastructure that neither EigenLayer nor Symbiotic offers directly. The honest weakness is demand rather than design: restaked capital and live network count both sit well below EigenLayer and Symbiotic, so the fee market for operators is thinner, and accepting stablecoins and long-tail ERC-20s as security collateral raises the question of what that collateral is actually worth in the scenario where it needs to be slashed and liquidated.
Solayer is the most architecturally distinctive entry here because it treats restaking and execution as the same problem. SOL and Solana LSTs are restaked to secure networks denominated in Solana's own economic security rather than Ethereum's, and the InfiniSVM work pushes throughput with hardware acceleration, including FPGA-assisted paths, rather than leaving performance to the consuming network. For a Solana-native service, that removes the awkwardness of renting security from a different chain's validator set and bridging the assumptions across. sUSD and the payments surface extended it beyond pure infrastructure. The honest weakness is scope and maturity: the addressable market is Solana-only, restaked capital is a fraction of EigenLayer's, slashing is specified but has seen limited production exercise, and the hardware-acceleration thesis is a claim about future throughput rather than a property a reader can verify today. For Solana infrastructure needing Solana-denominated security, it is the natural choice.
Jito Restaking is included because it is the framework Solana networks reach for in practice, and that is a different claim from being the largest. Jito's position in Solana validator infrastructure — its MEV distribution client runs across a large share of stake — means its Node Consensus Network standard arrives with distribution already attached, which is the hardest thing for a restaking layer to manufacture. The tooling is clean, the NCN abstraction maps well onto how Solana services are actually built, and integrating does not require a network to adopt an unfamiliar operator model. The honest weakness is that restaking is adjacent to Jito's main business rather than central to it: restaked capital dedicated specifically to NCN security is modest, slashing is less developed than EigenLayer's or Babylon's, and the protocol's incentives are shaped primarily by liquid staking and MEV rather than by the restaking market. Useful standard, smaller security budget.
Mellow solves the problem Symbiotic's flexibility creates. If a network can set any slashing condition and any ERC-20 can be collateral, then a depositor choosing a vault is making a risk assessment they are almost certainly unqualified to make. Mellow inserts named curators — firms that publish a mandate, select networks and operators, and are accountable for the configuration — so the depositor chooses a curator and a strategy rather than a parameter set. That is a genuine product rather than a wrapper, and it is why a meaningful share of Symbiotic's restaked capital arrives through Mellow vaults. The honest weakness is that it adds a layer without removing the underlying risk: a depositor now carries curator risk on top of network and collateral risk, curator track records are short because the category is young, and the fee taken for curation compounds against yield that is already thin relative to the slashing exposure being accepted.
| Protocol | Collateral | Restaked TVL | Native token | Slashing | Score |
|---|---|---|---|---|---|
| EigenLayer | ETH, LSTs, selected ERC-20 | $10B+ | EIGEN | In production | 8.9 |
| Symbiotic | Any ERC-20, permissionless | $1B–$3B | None at review date | In production | 8.4 |
| Babylon | Native BTC | $3B–$6B | BABY | In production | 8.0 |
| Karak | LSTs, LRTs, stablecoins | $300M–$900M | KAR | Limited production | 7.4 |
| Solayer | SOL and Solana LSTs | $200M–$600M | LAYER | Limited production | 7.1 |
| Jito Restaking | JitoSOL, SOL, LSTs | $100M–$400M | JTO | Early production | 6.9 |
| Mellow | ETH, LSTs, curator-selected | $300M–$800M | MLW | Via Symbiotic | 6.6 |
EigenLayer wins this ranking on the two things that matter most in a security market: the deepest restaked capital and the largest set of networks actually paying for it. Symbiotic is the better answer when a network needs parameters the shared framework will not give it, and its vault layer is the cleaner architecture of the two. Babylon wins outright for Bitcoin-denominated security and deserves attention for a reason unrelated to size — it is the only design here that puts BTC to work with no bridge, wrapper or custodian in the trust path.
Below those three, the ranking is about fit rather than quality. Karak's collateral breadth and K2 staging layer are real advantages for a specific kind of team, offset by thinner demand. Solayer and Jito Restaking split the Solana market between architectural ambition and integration convenience, and a Solana service will usually find the second more useful than the first. Mellow is the only entry that is a risk-packaging product rather than a security layer, and it is addressing a real gap — at the cost of a fee and another counterparty.
The honest closing note is that restaking is a risk product and this ranking should be read that way. Slashing is in production on the leaders now, so a restaker is accepting a penalty that can actually be levied, usually for a yield spread of a few percent. Before committing capital, work out which networks your operator or curator has opted into, what the slashing conditions are, and what the collateral would be worth in the scenario where those conditions trigger. If you cannot answer those three questions, the yield is not compensating you for the risk you are holding.
Deeper dives on specific matchups from this ranking.
See how your project ranks against the leaders in AI search and crypto SEO. No credit card. Free tier on one domain.
Run free audit →