Prior to the conclusion of this quarter, ether.fi will sever the final architectural bond connecting its staking assets to the restaking platform EigenLayer. According to protocol documentation, less than 1% of assets remained restaked as of August, with EigenPod withdrawal permissions scheduled for deletion before the year ends.
When ether.fi debuted back in 2024, user deposits were automatically restaked via EigenLayer. By August, the enterprise excised restaking functions out of weETH, the iteration of its asset circulating and utilized as collateral across decentralized finance, converting it strictly into a standard liquid staking token. Individuals desiring continued restaking must manually select a distinct token powered by Symbiotic, a competing network.
Ether.fi Chief Executive Mike Silagadze stated that this choice stemmed entirely from risk management considerations. “There were no meaningful yield opportunities in restaking and there was some perceived risk from stakers, so we decided it made sense to exit,” he informed CoinDesk.
What restaking was meant to be
Staking involves locking up ETH to support Ethereum network security in exchange for generated yields. Restaking emerged from the concept that this exact locked ETH could perform dual duties: EigenLayer would lease out that security apparatus to external applications, such as data availability layers and oracles, which in turn compensated for the protection. Lenders would consequently collect a double return on a single capital pool. Liquid restaking tokens functioned on top of this architecture, supplying depositors with tradable receipts that could be liquidated or utilized as external collateral instead of forcing users to freeze their ETH indefinitely. weETH stood out as the largest among them.
EigenLayer peaked at a total value locked of $19.7 billion, and liquid restaking tokens surged by upwards of 1,000% throughout the initial six weeks of 2024. Nonetheless, the protocols purchasing this security never disbursed sufficient funds to cover both standard staking returns alongside an additional premium, meaning the secondary yield promised by restaking never materialized.
By September 8, DefiLlama’s restaking segment held $10.02 billion while producing a mere $99,977 in fees over the preceding week. Conversely, the liquid staking sector, managing $51.87 billion, generated $27.35 million. Calculated per secured dollar, conventional staking earns roughly 53 times greater revenue.
Two separate events subsequently eliminated what remained of restaking incentives. Incentive campaigns designed to subsidize deposits phased out throughout 2025, and slashing protocols went live in April 2025. Slashing introduces financial penalties that confiscate fractions of an operator’s staked ETH when misconduct occurs, such as going offline or validating contradictory messages. Consequently, restaking instantaneously introduced tangible, priced liabilities where risks were previously purely theoretical, lacking any supplementary yield to offset the hazard.
Excluding ether.fi, the remainder of the ecosystem remains relatively small. Renzo, Kelp, Swell, Puffer Finance, and Bedrock—representing the five largest remaining liquid restaking tokens—secured a combined gross profit of $953,350 during the second quarter of 2026. Three quarters prior, those same five entities accumulated $2.18 million. Puffer, which secured $23 million in funding, reported $21,590 for the quarter, while Swell registered $22,370.
The corresponding income statements further illuminate which business segments actually drove profitability, revealing it was never the restaking mechanism. Kelp’s ledger books EIGEN token rewards as $460,600 in gross revenue offset by $460,600 in cost of revenue, signifying these assets simply pass directly to depositors without retaining value for the protocol. Puffer and Swell account for staking rewards through an identical mechanism. Whatever profits these enterprises accumulated originated exclusively from foundational staking commissions collected beneath the restaking architecture.
The hack
On April 18, a malicious actor exploited Kelp’s cross-chain bridge—the mechanism facilitating token transfers across blockchains—and minted 116,500 rsETH within 46 minutes, representing roughly $293 million backed by zero underlying ETH. The perpetrator deposited these assets into Aave as collateral to borrow authentic ether against them. Approximately $6 billion departed Aave in subsequent days, exposing potential bad debts ranging from $123 million to $230 million. Throughout May, Aave overhauled its collateral listing frameworks to evaluate technical infrastructure and cybersecurity alongside simple price fluctuations.
Silagadze pushes back against interpreting this incident as a failure inherent to leverage. “The cause of the Kelp hack was poor security practices with respect to cross-chain, not related to leverage,” he explained. “The ether.fi Aave market has very conservative parameters and we have a strong commitment to security.”
He remains accurate in noting that EigenLayer itself maintained functionality. No validator assets faced slashing, no restaking framework malfunctioned, and the point of failure resided entirely within a bridge. Nevertheless, this exact dynamic amplified the damage inflicted upon liquid restaking tokens. Capital losses occurred within the wrapper—the tradable receipt layer positioned atop restaking—rather than within the restaking architecture itself. By April, holders of these assets assumed auxiliary software vulnerabilities without collecting any supplementary returns. The wrapper had ceased covering its operational costs.
Where the money went
Capital fleeing restaking did not exit crypto lending entirely; rather, funds shifted from ETH allocations into dollar-denominated assets.
Throughout 2024, the prevailing strategy involved staking ETH, restaking it, wrapping the position inside a liquid restaking token, borrowing against it to acquire additional assets, and stacking exposure to a singular token. By 2026, equivalent behaviors manifested via curated vaults. A curated vault represents a lending pool where an external entity, designated as a curator rather than the core lending protocol itself, dictates accepted assets and operational terms in exchange for fee distributions. Morpho, serving as the largest marketplace for this model, manages roughly $5.8 billion.
Similar to restaking models, the asset held by a depositor operates as a receipt whose underlying risk parameters are defined externally and accepted as collateral by a third party.
Curated vaults have already generated their proprietary parallel to the Kelp collapse. On November 4, 2025, Stream Finance reported roughly $93 million in losses while freezing withdrawal functions. Its xUSD token, a yield-generating stablecoin designed to maintain parity at $1, plummeted 77% within a single day. Curators had established Morpho vaults where participants supplied authentic stablecoins against xUSD collateral, utilizing borrowed stablecoins to purchase additional xUSD, thereby artificially inflating the asset far beyond its genuine backing. Those markets treated xUSD at a fixed $1 valuation instead of actual market pricing, meaning that when real market values dropped, automated liquidations meant to close positions failed to trigger. Analysts subsequently mapped approximately $285 million in debt exposures across various lending venues. A secondary dollar token, backed 65% by loans tied to Stream, collapsed nearly 98% prior to being unwound.
The departure from restaking left ether.fi confronting the fundamental inquiry facing every liquid restaking protocol: What constitutes the underlying enterprise once its foundational product stops generating revenue? Their resolution involved abandoning identity as a pure staking entity.
Ether.fi’s second act
Ether.fi currently operates a payment card enabling users to spend against crypto holdings without liquidating assets, a lending marketplace on the Ethereum layer-2 network Optimism, a suite of vaults, and positions itself as a crypto neobank. In August, the platform integrated tokenized equities, precious metals, and fiat currency rails. Silagadze estimates the addressable neobank industry generates roughly $300 billion annually, dwarfing decentralized finance by a factor of 300.
Payment card operations expanded their contribution to monthly revenue from 17% in January up to 46% by July. “Neobank revenue has fully replaced the revenue lost from restaking and lower ETH price,” Silagadze noted. “We are on track to increase revenue overall run rate this year by about 38%, while staking and restaking revenue has declined by 70%. Diversification of our revenue has been a huge success.”
DefiLlama metrics indicate a contrasting trajectory across a comparable duration, documenting ether.fi gross profits dropping 47% from $18.71 million in the third quarter of 2025 down to $9.99 million throughout the second quarter of 2026. Both assessments can coexist, given that gross revenue differs from gross profit and forward projections differ from trailing quarterly data, though ether.fi has not publicly disclosed the calculation methodology behind the cited 38% growth.
Card transaction commissions yielded $3.14 million in gross profit during the second quarter. EigenLayer restaking generated $2.87 million, surpassing core ETH staking alongside vault fees, borrowing, and management charges combined. According to DefiLlama ledger accounting, restaking stood as ether.fi’s second most lucrative revenue channel at the exact moment leadership elected to exit.
Silagadze challenges a specific metric integrated within DefiLlama statistics. Cashback incentives distributed to card users appear within DefiLlama ledgers as both $5.83 million in revenue and corresponding expenses, rendering zero net impact on profitability. “That was back when third-party partners were paying the cashback rewards,” he stated. “That’s no longer the case, so current revenue reporting doesn’t include cashback subsidy grants.” DefiLlama’s tracking adapter continues recording both entries across the most recent quarter.
He likewise emphasizes that community participants received ongoing notices as ether.fi executed operational shifts. Migrating cash vaults onto Aave superseded a proprietary debt management system and mitigated overall risk, remarking: “Users were notified multiple times and opted in to this change.”
The technology was never the problem
None of these developments imply technological failure. EigenDA, the data availability mechanism built by EigenLayer, operates on the mainnet processing 100 MB/s and retains its status as the leading service measured by secured capital. Symbiotic, the alternative network selected by ether.fi for ongoing restaking, has successfully integrated upwards of 50 networks.
The central question never concerned whether restaking functions technically. Instead, the uncertainty focused on whether it generates sufficient cash flow to support a viable commercial enterprise, and for enterprises whose entire value proposition relied on restaking, the response has been negative.
EigenLayer has likewise ceased marketing restaking as its core offering. Rebranded as EigenCloud, the platform currently promotes verifiable computing, allowing applications to mathematically verify off-chain computational integrity, utilizing restaked collateral as foundational infrastructure rather than the primary product for sale. Platform assets currently total $5.10 billion, down significantly from $22.06 billion in August 2025.
The broader industry has yet to reach consensus on whether a market contracting approximately 75% while continuing to secure identical underlying volumes represents a failure, or if the sector simply expanded four times larger than actual operational demand warranted. Silagadze declined to wait for an answer, successfully transitioning ether.fi away from restaking while the debate persisted.
Originally published at https://www.coindesk.com/business/2026/09/28/the-restaking-gold-rush-is-over-and-top-protocols-are-barely-making-a-profit.
