TechFlow Logo
Login/ Sign up
ETH Gas
Gwei
Fear
gas
As the restaking craze subsides, why is ether.fi bidding farewell to EigenLayer?

As the restaking craze subsides, why is ether.fi bidding farewell to EigenLayer?

2026.09.29
Share
ether.fi will sever its final ties with EigenLayer this quarter, leaving less than 1% of assets still participating in restaking.

Written by: Oliver Knight

Translated by: Chopper, Foresight News

By the end of this quarter, ether.fi will remove the last structural tie between its staking tokens and the restaking protocol EigenLayer. Withdrawal credentials for EigenPods will be phased out by year-end, leaving less than 1% of assets still participating in restaking.

When ether.fi launched in 2024, user deposits of ETH were automatically restaked on EigenLayer. In August this year, the project decoupled the restaking functionality from weETH. As ether.fi's most liquid staking token and widely used as collateral in DeFi, weETH now functions as a standard liquid staking token after the split. Users wishing to participate in restaking must opt-in separately to another token deployed on the competing platform Symbiotic.

In an interview with CoinDesk, ether.fi CEO Mike Silagadze stated that the decision was driven by risk considerations: "There are no longer substantial yield opportunities in the restaking sector, and stakers have become aware of the risks, so we chose to exit."

The Original Vision of Restaking

Staking involves locking up ETH to secure the Ethereum network and earn rewards in return. The concept behind restaking is that the same locked ETH can perform a second function: EigenLayer rents out this security capacity to other services such as oracles and data availability layers, which pay fees for it. Depositors thus earn dual yields from the same capital. Liquid restaking tokens are built on this mechanism; depositors receive tradable vouchers, avoiding lock-up periods. These tokens can be sold or used as collateral elsewhere, with weETH being the largest among them.

EigenLayer's TVL once peaked at $19.7 billion, and the market cap of liquid restaking tokens surged more than tenfold in the first six weeks of 2024. However, the fees paid by demanders purchasing security services have consistently failed to cover basic staking yields and risk premiums, meaning the promised secondary layer of returns from restaking has never materialized.

As of September 8, DefiLlama data showed that total TVL in the restaking sector stood at $10.02 billion, generating only $99,977 in fees over the past week. In contrast, the liquid staking sector had $51.87 billion in TVL and generated $27.35 million in weekly fees. This implies that for assets of equal size, traditional staking generates approximately 53 times more yield than restaking.

Restaking protocols barely generate yield

Two developments further eroded incentives to participate in restaking: points programs subsidizing deposits gradually ended throughout 2025; and the slashing mechanism officially went live in April 2025. Slashing penalizes validators by confiscating a portion of their staked ETH for infractions such as going offline or signing conflicting messages. This shifted restaking from purely theoretical risk to tangible losses, with no extra yield to compensate for the added exposure.

Excluding ether.fi, the remaining projects in the sector are relatively small. The top five liquid restaking protocols—Renzo, Kelp, Swell, Puffer Finance, and Bedrock—generated a combined gross profit of only $953,300 in Q2 2026. Three quarters prior, their combined gross profit stood at $2.18 million. For instance, Puffer, which raised $23 million, reported a quarterly gross profit of just $21,590; Swell reported $22,370.

DefiLlama data also reveals that the truly profitable part of these operations has never been restaking itself. Taking Kelp as an example, EIGEN token rewards recorded $460,600 in gross revenue, with identical costs of $460,600—the rewards were fully distributed to depositors, leaving zero for the protocol. Puffer and Swell followed the exact same model. All profits for these protocols come solely from underlying basic staking fees.

The Hacking Incident

On April 18, 2026, attackers exploited a vulnerability in the KelpDAO cross-chain bridge to mint 116,500 rsETH out of thin air within 46 minutes, worth approximately $293 million without any underlying ETH backing. The attackers deposited the counterfeit tokens into Aave as collateral and borrowed real ETH. Within days, roughly $6 billion in funds withdrew from Aave, resulting in potential bad debt ranging from $123 million to $230 million. In May, Aave announced a comprehensive overhaul of collateral and listing standards, simultaneously reviewing projects' cybersecurity, technical architecture, and price volatility risks.

Silagadze does not view the incident as a failure of leverage mechanisms. He stated, "The root cause of the theft from Kelp was a flaw in the cross-chain bridge's security, unrelated to leverage. ether.fi maintains extremely conservative market parameters on Aave, reflecting our top priority on safety."

Objectively speaking, EigenLayer itself experienced no failures or slashing events, meaning the core restaking mechanism remained intact; the weakness lay solely in the cross-chain bridge. Yet this precisely explains why the hack severely damaged the liquid restaking sector: capital losses occurred at the wrapper token layer (i.e., tradable vouchers) above restaking, not within the underlying protocol. By April, holding these tokens meant assuming an extra layer of smart contract attack risk without receiving corresponding additional yield. The yield from wrapper tokens was already insufficient to cover their own inherent risks.

Where Did the Money Go?

Capital withdrawing from the restaking sector did not leave the crypto lending market; it merely shifted from ETH-denominated assets to USD-backed assets.

The mainstream strategy in 2024 involved staking ETH, restaking it, wrapping it into a liquid restaking token, using it as collateral for borrowing, and then leveraged again to stack exposure to a single asset. By 2026, similar capital maneuvers shifted toward curated vaults. Curated vaults are lending pools where external entities (vault managers), rather than the lending protocol itself, decide which assets are accepted and set specific terms, taking a cut in return. Morpho is the largest platform in this segment, with approximately $5.8 billion in TVL.

Similar to restaking logic, depositors only receive a voucher, while risk parameters are determined by third parties. These vouchers can then be accepted as collateral by other platforms.

Curated vaults have also experienced crashes similar to Kelp's. On November 4, 2025, Stream Finance disclosed losses of approximately $93 million and halted withdrawals. Its yield-generating USD stablecoin, xUSD, aimed to peg at $1 but plummeted 77% in a single day. Vault managers deployed vaults on Morpho; depositors would deposit real stablecoins to obtain xUSD, borrow stablecoins to purchase more xUSD, artificially inflating the token's price far beyond its actual asset backing. The market consistently priced xUSD at a fixed $1 rather than its true market value, causing automatic liquidations that should have triggered during the crash to fail. Subsequent analysis estimated related debt exposure across lending platforms at around $285 million. Another USD-denominated token (backed by 65% funding from Stream loans) crashed roughly 98% before being liquidated.

ether.fi's exit from restaking also highlights a challenge faced by all liquid restaking protocols: when the very business model they built upon ceases to be profitable, what comes next? ether.fi's answer is clear: no longer operating solely as a staking service provider.

ether.fi's Second Growth Curve

ether.fi has since launched a crypto consumer card allowing users to spend directly without selling crypto assets; it has also deployed a lending market on the Ethereum Layer 2 network Optimism and built a series of vaults. In August this year, ether.fi further expanded into neo-banking by launching tokenized stocks, metal assets, and composite asset loans built on Aave.

Silagadze estimates the neo-banking sector generates an annual revenue of roughly $300 billion, making it 300 times larger than the current DeFi market. Revenue contribution from the consumer card rose from 17% in January to 46% in July. He remarked, "Neo-banking revenue has completely offset the shortfall caused by contracting restaking activity and falling ETH prices. We project overall annualized revenue growth of approximately 38% this year, while staking and restaking income declined by 70%. Diversification has been a massive success."

Data from DefiLlama presents a contrasting trend: ether.fi's gross profit fell from $18.71 million in Q3 2025 to $9.99 million in Q2 2026, a drop of 47%. The two sets of figures do not necessarily conflict: revenue differs from gross profit, and annualized projections differ from historical quarterly reports. However, ether.fi has not disclosed the methodology behind its claimed 38% growth rate.

Gross Profit by Business Segment for ether.fi in Q2 2026

In Q2, credit card transaction fees generated $3.14 million in gross profit. The EigenLayer restaking segment contributed $2.87 million, surpassing both the base ETH staking business and the combined totals of vaults, lending, and management fees. According to DefiLlama's statistics, at the time of deciding to exit the sector, restaking remained ether.fi's second-largest profitable business line.

Silagadze disputes DefiLlama's calculation methodology. Cashback rewards are recorded in DefiLlama's metrics as both $5.83 million in revenue and $5.83 million in cost, contributing zero to actual profit. He explained, "These rewards were initially subsidized by third-party partners. Rules have since changed, and current revenue tracking no longer includes these subsidies. However, DefiLlama still applied the old accounting method for recent quarters." He also noted that migrating the cash vault to Aave replaced their custom debt management system, thereby reducing risk. "We notified users multiple times, and they voluntarily opted into this change."

The Technology Itself Was Never Failing

This does not imply that restaking technology itself has failed. EigenLayer's data availability service, EigenDA, achieves mainnet throughput of up to 100MB/s, maintaining the largest secured asset footprint in the industry. Meanwhile, Symbiotic, the platform ether.fi migrated its restaking operations to, has integrated over 50 networks.

The core contradiction lies not in whether restaking technology can function technically, but whether it can generate sufficient revenue to sustain its business model. For protocols that rely on restaking as their sole core product, the answer is no.

EigenLayer has similarly stopped positioning restaking as its primary offering. Rebranding itself as EigenCloud, it now focuses on verifiable computation, helping applications prove the authenticity of off-chain computational results. Restaking collateral has transitioned into foundational infrastructure rather than a marketed core product. Its TVL has contracted significantly from a peak of $220.6 billion in August 2025 down to $5.1 billion.

The industry remains undecided: TVL shrank by approximately 75%, yet the volume of secured services did not diminish. Is this considered a sector failure, or simply proof that the previous scale was four times larger than actual demand? Silagadze did not wait for this debate to conclude, instead opting to lead ether.fi out of the restaking space ahead of schedule.