EigenLayer: up to 8.4 billion in TVL, 7 in fees

The services secured by EigenLayer paid $607,375 in fees over the last 30 days. The capital protecting them ranges between $5.2 and $8.4 billion. This disproportion —billions in collateral versus less than a million in monthly revenue— is the entire story of restaking in 2026. EigenLayer (rebranded as EigenCloud after its 2025 product pivot) built the "shared security" market: a protocol where ETH already deposited in Ethereum is re-committed (restaking) to back other services in exchange for a fee. The security supply arrived. The demand meant to pay for it did not. This article crosses both data series in the same time window —secured TVL against real fees from AVS (Actively Validated Services, the clients renting that security)— and explains why the yield a restaker earns today is primarily EIGEN token emissions, something the protocol's own governance admits in writing in its ELIP-12 proposal.

What truly secures a TVL of 8.4 billion if almost no one pays the premium?

The TVL (Total Value Locked) of EigenLayer depends on how it is measured, and the discrepancy is massive. DefiLlama reports 5.172 million $ as of July 23, 2026, which is 77% below the all-time high of 22.060 million $ that the platform recorded in August 2025. Aggregators measuring raw restaked ETH —4.36 million ETH, placing EigenLayer's market share at 93.9% of the entire restaking market— currently equate to approximately 8.400 million $ at the current ETH price. The difference is methodological: what is considered an active deposit versus what is discounted due to double-counting of liquid restaking tokens.

For the thesis, the choice of denominator is secondary. Whether with 5,200 million or 8,400 million, the fee flow coming from the secured services is tiny. Insurance makes sense when someone pays the premium; here, the capital is sitting idle, ready to respond if a service fails, but the services are barely buying coverage. The drop in TVL from 22,000 million a year ago —passing through 12,600 million $ in January and 8,860 million $ in April 2026— is in itself a signal: part of the capital that entered due to yield expectations is already leaving upon realizing where that yield was coming from. EigenLayer is currently a security supply market without sufficient buyers: billions in collateral against 607,375 dollars in fees over 30 days.

What is an AVS and why is it supposed to pay for security?

An AVS is any service—a price oracle, a data availability layer, a cross-chain bridge, a sequencer network—that, instead of recruiting, coordinating, and paying its own set of validators from scratch, rents the economic security already deposited in EigenLayer. The 4.36 million ETH restaked as of July 10, 2026, serves as the collateral: if the AVS behaves incorrectly and the operators validating it cheat, that capital can be slashed. In exchange for providing the collateral, the restaker expects to collect a fee paid by the AVS.

The foundational thesis of 2023-2024 was elegant: starting a new protocol is extremely expensive because it needs its own security, and EigenLayer allowed "renting" Ethereum's security immediately. Two years later, the technical scaffolding works—there are operators, there are AVS in production, slashing is active—but the market has only developed on one side. There is plenty of cheap security supply and very little demand willing to pay for it. To understand why, one must look at what these services actually charge.

A useful mental parallel is the cloud. Amazon Web Services or Google Cloud rent computing capacity to those who need it, and their value is measured by what their clients bill, far more than by the size of their data centers. EigenLayer aspires to be that cloud, but for security: a market where AVS are the clients paying to "rent" validators. The pivot to the EigenCloud brand in June 2025—accompanied by a $70 million round led by a16z, with EigenAI, EigenCompute, and EigenDA as product lines—makes this ambition explicit. The difference with the real cloud is that AWS bills tens of billions to clients who have no cheap alternative; here, potential clients do have one, and in many cases, they choose it.

How much do EigenLayer AVS actually earn in 2026?

This is where crossing the data series moves beyond rhetoric. These are the fees that EigenLayer AVS distributed through the protocol, according to DefiLlama accounting as of July 10, 2026, alongside the TVL that same capital secures:

Annualizing the pace of the last 30 days —$607,375 multiplied by twelve— the real fee flow from AVS is around $7.3 million per year. Based on DefiLlama's TVL, that represents a fee yield of 0.14% per year; based on the gross count of approximately $8.4 billion, it drops to 0.09%. Even taking the most generous figure possible —the $43.8 million from the last twelve months, which includes past periods of higher activity— the fee yield on TVL remains below 1%.

It is worth grounding the scale. A protocol with derivatives like Hyperliquid generates more fees in a single day (~$1.5 million, DefiLlama as of July 23, 2026) than EigenLayer does in two and a half months ($607,375 in 30 days); the comparison is exactly the point. When an infrastructure moves billions in collateral but invoices less than a million per month, the collateral reflects an appetite for a subsidy rather than real demand for a service. The daily series confirms this and adds a decisive nuance: the two largest fee peaks were recorded on July 1 and 2, 2026 —$66,294 and $73,128— precisely on the day of the EIGEN unlock and the following day. Between July 3 and 6, fees returned to their baseline, from $652 to $1,918 per day, with a second minor spike arriving on the 7th and 8th —$51,972 and $44,295— and a third isolated peak of $57,188 on July 22, while the daily base for the middle of the month sat between $0 and $2,800. The fact that the maximum AVS activity coincides with the token unlock, rather than any new adoption, reinforces the thesis: the movement is driven by the token, not the business. In terms of revenue, EigenLayer earns per month what a small local business invoices in a few weeks, far below what its TVL would suggest.

The revenue captured by the protocol as of July 23, 2026, is exactly zero. The fees paid by AVS are passed in full to operators and restakers; neither the protocol nor the EIGEN token retains a single cent. There is no active "fee switch." This is the same debate we analyzed in the DeFi fee switch and buyback comparison, and precisely what ELIP-12 attempts to change.

Why hasn't EigenDA, the flagship AVS, won the data availability market?

EigenDA is EigenLayer's data availability layer—the service that stores and ensures that data from an L2 (Layer 2 network on Ethereum) is available for anyone who wants to verify it—and it is the ecosystem's flagship AVS. On paper, it is competitive: for a high-volume rollup (an L2 that bundles many transactions off-chain and settles them on Ethereum) publishing 100 MB per day, EigenDA costs roughly $730 annually compared to $12,775 for Celestia, its modular rival, while offering higher throughput per second.

Despite this price, it has not won the market. Celestia maintains nearly 50% of the data availability market share as of March 2026, and the largest L2s continue to publish their data directly to the Ethereum mainnet via blobs—the mechanism that drastically lowered that cost after the network's latest updates—rather than outsourcing it. EigenDA adoption exists but is niche: Mantle, an L2 on Ethereum, is the most cited example. The reason is easy to name and hard to solve: EigenDA's security relies on data availability committees, while Celestia uses data availability sampling, which many teams consider more robust, and the Ethereum mainnet offers maximum security without intermediaries. The market's verdict is numerical: the cheapest provider—$730 annually versus Celestia's $12,775—is not the leader.

Is restaking yield real or is it EIGEN emissions?

If real fees yield 0.21% annually, the obvious question is why anyone deposits billions. The answer lies in the token. The bulk of the yield a restaker perceives today does not come from AVS fees, but from EIGEN emissions: newly created tokens that the protocol distributes as an incentive to attract capital. It is yield paid with the token's own inflation, financed by dilution rather than service usage.

This distinction is the backbone of all the real yield analysis we perform on the site, from the DeFi real revenue ranking to the contrast with protocols where revenue has been measured, such as in the Hyperliquid revenue fundamentals. The emission subsidy mechanic works as long as the token price holds and capital believes real fees will follow. The risk appears when that expectation cools and the token paying the yield begins to dilute due to scheduled unlocks, as occurred with the July 1, 2026, unlock.

What happened on July 1, 2026, with the EIGEN unlock?

On July 1, 2026, 36.82 million EIGEN tokens (~$8.7 million at that day's price) were released through a vesting cliff—a sudden unlock of allocations to early contributors. That week, the overall unlock market totaled around $73 million across all protocols—led by Sui and Ethena (SUI and ENA) tokens—not just EIGEN. The unlock coincided with an EIGEN price rally of around 14%, driven by on-chain capital inflows and high long-to-short ratios in derivatives (1.29 on OKX, 1.53 on Binance), a sign of a market betting upward just before the major supply pressure.

Price context is what sets the stage. As of July 23, 2026, EIGEN is trading at $0.23, with a circulating market cap of $172.0 million based on 741.2 million tokens in circulation and a fully diluted valuation of $423.6 million. It is down 95.9% from its all-time high of $5.65, reached in December 2024. A yield-subsidizing token trading at less than one-twentieth of its peak, while releasing monthly unlocks worth tens of millions, is an incentive engine running on empty. That is the problem that governance acknowledged in writing.

What does ELIP-12 try to fix and why is it not a guaranteed solution?

ELIP-12 is the governance proposal published by the Eigen Foundation in December 2025 that admits the problem without euphemisms: most restaking yield today comes from EIGEN emissions, not AVS fees from real usage. To correct this, it proposes three components:

• A 20% fee (the so-called fee switch) on AVS rewards that are subsidized by EIGEN incentives, diverted to a fee contract that can be used to buy back the token.
• 100% of EigenCloud revenue—from infrastructure products EigenAI, EigenCompute, and EigenDA—destined for that same buyback contract as those revenues grow.
• An Incentives Committee, planned without a confirmed date by the Foundation, tasked with directing emissions toward AVS that actually generate fees, rather than distributing them indiscriminately.

The logic is to close the loop: ensuring the token stops being just a subsidy and starts capturing value from real activity. However, the proposal does not create demand; it only redirects what already exists. A 20% fee on 7.3 million in annual fees amounts to 1.5 million per year for buybacks: relative to a market cap of $172.0 million, this is a marginal impact as long as AVS do not pay significantly more. Furthermore, 100% of EigenCloud revenue depends on EigenAI, EigenCompute, and EigenDA generating sales — precisely the area where EigenDA has yet to take off. ELIP-12 is a genuine and well-designed attempt to link rewards to real income. Its success depends on a variable that no treasury reform can control: someone actually buying the security.

How to read any restaking TVL figure from now on?

The framework left by this case is a single question, applicable to any restaking protocol—including the competitors we cover in EigenLayer vs. Symbiotic—and to any TVL headline from here on: of this secured capital, how much generates real fees today and how much is just a deposit waiting for emissions? The TVL figure alone says nothing about the health of a security market; it can reflect genuine demand or capital hunting for a depleting subsidy.

With EigenLayer, the arithmetic of July 2026 is what it is: billions in collateral, about 7.3 million in annualized real fees, zero revenue captured by the protocol, and a token that subsidizes the difference while trading 95.9% below its highs. The restaking infrastructure is a genuine technical achievement and ELIP-12 targets the right lever. However, a security market without enough buyers is not fixed by offering more security: it is fixed when AVSs have their own business to protect. Until the fees in the table above grow by an order of magnitude, any reading of restaking TVL should start by dividing the figure by what that capital actually earns.

Sources and links: DefiLlama — EigenCloud TVL and Fees · EigenFoundation — A New Era of EIGEN Incentives (ELIP-12) · CoinDesk — Eigen Foundation plans bigger rewards for active users · AInvest — July 1 cliff unlock · CoinGecko — EIGEN price and market cap · DA Layers — Celestia vs EigenDA 2026