Restaking

Restaking lets stakers earn additional yield on already-staked tokens by committing them to secure other networks. EigenLayer is the reference implementation: restaked ETH serves as cryptoeconomic security for protocols beyond Ethereum, in exchange for protocol fees and rewards. The modules it secures are Autonomous Verifiable Services (AVS), renamed from Actively Validated Services in February 2025, and the current framing is deliberately broad: any off-chain service that can be verified on-chain, including rollup services, co-processors, cryptography services and zk proof services. The longer list of consensus protocols, data availability layers, virtual machines, keeper networks, oracle networks, bridges, threshold cryptography schemes and trusted execution environments comes from the 2023 whitepaper and describes that era rather than the current design. Operators run the AVS software and restakers delegate to them, to one operator at a time and all or nothing. Rather than each service bootstrapping its own validator set, EigenLayer aggregates Ethereum's security across all of them, a design its whitepaper calls pooled security, with rollups consuming the resulting services. EigenDA, built by EigenLabs, was the first AVS scheduled to launch on it. Eligible collateral is native ETH and liquid staking tokens such as Lido's stETH, plus EIGEN and other ERC-20 tokens, so the staked-a-second-time description fits the native and LST routes rather than every deposit, and an LST route is frequently two layers deep before the restaking even begins.

That layering is the part a supply analyst should care about. Staked tokens sit outside the tradeable float, but liquid staking hands the holder a transferable claim on them, and restaking then pledges that claim again. Tokenomist's Ethereum Shanghai unlock dashboard put staked ETH above 20 million, roughly 17% of total supply, in mid 2023, and about 41.9 million ETH is staked as of August 2026, with liquid staking derivatives a major route in. Building a derivative on top of a derivative concentrates a specific failure: if trust in one protocol in the stack breaks and holders redeem back toward the original asset, the derivative can trade at a discount, as stETH did before unstaking became possible with the Shapella upgrade on 12 April 2023, which paired Shanghai on the execution layer with Capella on the consensus layer. stETH bottomed near 0.934 ETH on 19 June 2022, about 6.6% below par, a secondary-market discount rather than a shortfall of backing. Slashing exposure, by contrast, is not pooled. Since slashing went live under ELIP-002 on 17 April 2025, the Operator rather than the restaker allocates a proportion of delegated stake to a specific Operator Set, and that Unique Stake is solely slashable by the AVS that created the set, so an AVS carries no exposure to other AVSs or their slashings. Conditions are AVS-defined and need not be objectively attributable. In native restaking a slash does reach the same ETH that secures Ethereum, which remains permanently locked in the EigenPod contracts, but only the allocated proportion, and it does not trigger an Ethereum consensus-layer slashing. Redistribution of slashed funds exists as an opt-in path under ELIP-006, shipped in core contracts v1.5.0 on 7 July 2025, with native ETH and EIGEN not yet eligible and slashed funds otherwise burned.

Restaking also produced a cohort of tokens with their own vesting schedules, and those schedules do not care whether the narrative still works. PUFFER released 12.40% of its circulating supply on February 11, 2026 to Investors, Early Contributors and Advisors, an unlock worth only $654K but the largest in that week by proportion, landing while the token printed new all-time lows after the restaking narrative faded. Renzo's REZ was projected in December 2025 to release 22.34% of its total supply, about $15.21 million, over the following six months. $BB, in the BTC restaking segment, took a December 2025 cliff unlock that expanded supply by 14.86%. Narrative decay and supply expansion tend to arrive together, because the schedules were written when the narrative was strongest.

For a supply model, restaking changes what staked capital is doing, not how much of it exists. TVL in a restaking protocol is not float and should not be netted against circulating supply: deposits are user assets that can be withdrawn, while the protocol's own token follows its published schedule. Much of the early TVL was also bought rather than earned. EigenLayer and Ether.fi both ran points programs that awarded credit based on the size and duration of a deposit, with EigenLayer's points redeemable for an airdrop later, which means part of the deposit base was renting yield rather than committing to it. Ether.fi has since shifted toward revenue-funded buybacks, routing 100% of eETH withdrawal-fee revenue into weekly purchases on roughly $52.6M of annualised revenue as of mid-2026. Read the deposit number and the token schedule as two separate things.