Skip to content
Financial Context Wiki

Correlated slashing risk in restaking

Restaking can expose one portfolio to several slashable services. Learn how shared operators, software, infrastructure, and exit delays can turn separate allocations into correlated losses.

Updated

For educational purposes only; not investment advice. Restaking can cause partial or total loss of principal.

Direct answer

Correlated slashing risk is the possibility that one underlying failure causes losses across several restaking exposures at once. Shared operators, signing systems, client software, cloud regions, oracles, governance, and smart contracts can make apparently separate services fail together.

Restaking reuses economic stake to secure additional services, often called actively validated services (AVSs). Rewards may accumulate, but so can protocol, operator, smart-contract, liquidity, and exit risks. The advertised yield is not a reliable measure of the maximum loss.

The amount that can be slashed depends on the live contracts, the stake allocated to each operator set, each service’s slashing policy, and any still-slashable withdrawal or deallocation window. It must not be inferred from a generic percentage or from past incident counts.

How it works

In EigenLayer’s current unique-stake model, an operator allocates portions of delegated stake to operator sets. Allocations for one strategy cannot exceed 100% in total, and the same allocated portion cannot be unique stake for two operator sets. A service may slash the portion allocated to its operator set according to its own rules.

That accounting limits double-use of the same unique stake, but it does not make losses independent. One faulty deployment can affect several separately allocated portions, many operators can share the same dependency, and native ETH validators can also face Ethereum consensus penalties or slashing.

Net expected return = expected rewards - fees - opportunity cost - expected slashing and liquidity losses

Estimate loss by scenario, not by adding headline yields. Review at least these items:

  • Record the strategy, operator, operator set, allocated percentage, and slashable amount for every service.
  • Read each slashable condition, evidence standard, authority, dispute or veto process, and maximum slash.
  • Map shared clients, keys, signers, cloud regions, oracles, upgrade administrators, and governance participants.
  • Check withdrawal and deallocation delays, because queued shares may remain slashable until the applicable delay ends.

Example

Suppose 100 ETH is delegated to one operator: 40% is allocated to operator set A, 30% to operator set B, and 30% remains unallocated. If one shared software failure causes A to slash 50% of its allocation and B to slash 100% of its allocation, the restaking loss is 20 ETH + 30 ETH = 50 ETH. This is an illustration, not a universal formula; contract rules, later slashes, Ethereum penalties, and an LRT market discount can change the realized loss.

Risks

  • Hidden concentration: different operator names may still use the same client, signer, data center, or cloud region.
  • Rule and governance risk: a service may define broad slashing conditions or rely on a committee, upgrade key, or dispute process that fails.
  • Exit-tail risk: requesting withdrawal or deallocation may not end slashability immediately.
  • Liquidity-layer loss: an LRT can trade below the value of its underlying assets during stress, adding market loss before protocol accounting settles.

Common misconceptions

Myth 1: More AVSs automatically diversify risk

They diversify only when their failure sources and slashable allocations are genuinely independent.

Myth 2: No past slashing means the risk is low

A short or quiet history may contain no relevant stress event. Contract permissions and failure scenarios matter more than a zero-incident count.

Myth 3: An insurance fund guarantees repayment

Coverage depends on fund size, eligible events, exclusions, claim priority, and governance. A promise of insurance is not the same as funded, enforceable coverage.

Myth 4: Exiting ends liability immediately

Queued withdrawals or deallocations may remain slashable during protocol-defined delays, and an LRT holder may also face liquidity and redemption delays.

Sources

Navigation

Search the wiki...